Poster viewing session, complimentary drinks served

New formulation coatings with antimicrobial additives- dispersions of nanoparticles such as titanium dioxide, silica- can be an effective means to minimise the rapid spread of diseases in areas of dense populations (hospitals, schools, in public transport) and difficult-to-clean surfaces. This study was performed with titanium dioxide dispersions (10% w:w in water at pH≈3 - stability shown via zeta potential measurements) used as antimicrobial coating additives with the objectives to determine the kinetics and mechanisms of deagglomeration, dispersion fineness and provide guidelines for process design. Process intensification approach was taken using a batch rotor-stator with different mixer heads and an ultrasonic processor. The dominant mechanism of deagglomeration was erosion and dispersion fineness was defined by the mean aggregate size (~120 nm) regardless of the power input and dispersion method. Deagglomeration kinetics on the other hand was enhanced by increasing the power input. The effect was more prominent with the ultrasonicator and overall, kinetics were significantly faster allowing complete deagglomeration. Detailed quantitative analysis of kinetics will be presented at the Conference, also including comparisons to hydrophilic silica dispersions and agglomerate strength values. The rotor-stator alone could ascertain dispersion homogeneity; with the ultrasonicator an impeller was also required. Final dispersions demonstrated long-term stability. The dispersions were Newtonian with a low viscosity, in contrast to a more complex rheology noted with hydrophilic silica dispersions at a comparable concentration. We will report on final product/coating properties formed with pre-dispersions, partially and fully deagglomerated additives. Finally, recommendations for design and scale-up will be provided.

During the last decades the process intensification in the field of bioprocess and pharmaceutical engineering has made significant advances. Nevertheless, the demand of process intensification to develop and scale-up mammalian cell culture processes with high cell densities is still present. This work focuses on tests of new stirrer types for intensified oxygen mass transfer in pilot scale stirred tank reactors and the scale up to 15 000 L industrial scale. Improvements are compared to the commonly used combination of a Rushton turbine and a pitched blade stirrer. On pilot scale a significant improvement of the mass transfer performance was achieved for the combination of Rushton turbines with Combijet-like stirrers. The promising combinations have been characterized further in the transparent industrial sized stirred tank reactor at the Institute of Multiphase Flows in Hamburg. Significant difference in mass transfer performance and dependency on power input were observed compared to the investigations at pilot plant scale. This presentation aims to discuss the conducted scaleup study and the scale dependent parameters that need to be considered.

Particle settling and transport in slurries significantly affect the efficiency and energy consumption of slurry-based operations in industries such as energy, mineral processing, and wastewater treatment. In many industrial slurries, the carrier fluid exhibits a finite yield stress, yet the conditions under which particle wake structure changes with increasing inertia in these fluids remain incompletely characterized. In this study, direct numerical simulations are performed to investigate the gravity-driven settling dynamics of one and two spherical particles in Newtonian and Bingham fluids using the lattice-Boltzmann method coupled with an immersed boundary approach in a fully periodic domain. The numerical framework is first validated for Newtonian fluids against established numerical benchmark data over a broad range of particle Reynolds numbers, from creeping flow (Re < 1) to transitional regimes (200 < Re < 2000). The approach is then extended to Bingham fluids to capture yield stress effects over Bingham numbers of 0.8 to 8. These predictions are compared against the available numerical and experimental data. This work demonstrates that the yield surface fundamentally modifies particle wake structure and drag behavior and alters the wake transitions relative to Newtonian settling. These findings provide mechanistic insight for developing predictive models of particle-laden yield stress flows.

We investigate mixing enhancement in Taylor-Couette flow using vertically asymmetric rough walls through experiments and direct numerical simulations. The inner cylinder features triangular ribs (heights δ = 0.2d, where dd is the gap width), creating direction-dependent shear: clockwise rotation exposes the steeper slope, counter-clockwise the gentler slope. Results show that asymmetric roughness significantly enhances momentum transfer compared to smooth walls. At high Taylor numbers (Ta up to 2.39×10⁷), clockwise rotation yields up to 15% greater torque—indicating enhanced mixing—than counter-clockwise rotation. DNS reveals three synergistic mechanisms: (i) increased convective angular velocity flux from stronger wall-bulk coupling; (ii) thinner boundary layers and elevated Reynolds stress indicating intensified turbulence; and (iii) dominant pressure force contribution on rough surfaces, maximized for clockwise rotation. At low Ta, directionality has negligible effect due to viscous dominance. These findings establish vertically asymmetric roughness as an effective passive strategy for tuning mixing in wall-bounded turbulence, with implications for chemical reactors and rotating machinery. The combined approach validates mechanistic insights for optimizing mixing through tailored wall topography.

Emulsions stabilized by solids present many opportunities as new mineral separation processes. Their effectiveness is based on the fundamental mechanism of particle affinity: certain mineral particles preferentially attach to either the oil or water phase, which helps stabilize the emulsion. In this context, wettability refers to how easily a liquid spreads across or adheres to a solid surface. It is quantified by the contact angle, which is the angle formed between a droplet of liquid and the surface of a solid (angle>90° is more hydrophobic; contact angle < 90° expresses hydrophilicity). Quantifying wettability for fine particles is challenging due to their non-planar shapes and rough surfaces. To overcome this issue, we developed an innovative technique for measuring these properties. Our study addresses the challenge of selectively separating rare earth element (REE) minerals, such as bastnaesite and monazite, from gangue minerals like dolomite and calcite. We use a solid stabilized emulsification (SSE) process. In this process, particles with subtle differences in physicochemical surface properties—especially wettability as indicated by their contact angle with water—are separated. The valuable REE, which have higher contact angles and are thus more hydrophobic, preferentially attach to oil droplets. In contrast, the more hydrophilic carbonate minerals (contact < 48°) remain in the aqueous phase. Monazite and bastnaesite demonstrated a strong affinity for the oil, resulting in a 68% REE recovery and an enrichment ratio of 2.9 in a single-stage process. These results demonstrate that SSE is a promising approach for the beneficiation of fine REE.

The modeling of solvent extraction processes using coupled Computational Fluid Dynamics (CFD) and Population Balance Modeling (PBM) represents a robust solution for industrial applications. In this context, the simplified zero-dimensional (0D) PBM informed by single-phase CFD simulations can significantly reduce computational cost while providing results well suited to a wide range of applications. This approach accounts for the full turbulent energy spectrum by computing the second-order longitudinal structure function [1]. The objective of this work is to evaluate the impact of the CFD strategies (spatial and mesh resolutions, rotation modeling strategies - namely the Multi Reference Frame (MRF) and Sliding Mesh (SM) methods - and turbulence modeling approaches (standard k–ε model and Detached Eddy Simulation (DES) coupled with the realizable k–ε model) on the breakage frequency, Γ , and the resulting Sauter mean diameter, d32, within the reduced PBM using the modified Coulaloglou and Tavlarides breakage kernel [2]. Moreover, we investigate the evolution of the Taylor microscale and the Integral macroscale for each CFD strategy, relying on the findings of Escudié and Wu [3,4]. The results highlighted the domains beyond which k/ε has converged, leading to a maximum deviation of 5 % on the d32 estimation between the converged simulations. 1. Castellano, S. et al., Chemical Engineering Journal 374, 1420–1432 (2019) 2. Coulaloglou and Tavlarides, Chem. Eng. Sci. 32, 1289–1297 (1977). 3. Escudié, R. & Liné, AIChE Journal 49, 585–603 (2003). 4. Wu, H. & Patterson, G. K., Chem. Eng. Sci. 44, 2207–2221 (1989).

In many chemical and pharmaceutical processes, liquid–liquid dispersions are performed in batch stirred tanks. While stirred tanks can be suitable, energy dissipation is concentrated near the impeller, making uniform dispersion and controlled drop size slow and difficult to scale. By contrast, a properly sized in-line static mixer in a continuous pipeline can produce uniform dispersion in a single pass. Motivated by these differences, we evaluated static mixer integration in continuous flow loops at two scales to develop a mixing-energy-based scale-up approach. In prior work we studied two mixing tanks with different geometries and impellers and explored correlations of Sauter mean chord length with power and mixing energy. Those results showed that characterizing process dynamics via the mixing energy collapses data across diverse geometries and that a promising correlation exists between mixing energy and the Sauter mean chord length at the just-dispersed condition. Experiments in this new work used a 12 mm bench loop (2.5 m, 0.5 L) and a 50.8 mm industrial loop (24 m, 53 L) equipped with a commercial static mixer. Chord length distributions were monitored continuously with an FBRM probe while pressure losses across pipe sections and the static mixer were recorded. The data were used to compute component-level energy dissipation rates and Kolmogorov length scales. Tracking the evolution of chord length distribution over cycles together with component-level energy dissipation can help predict and replicate droplet distributions during pipeline scale up. Statistical analysis of these descriptors will support explicit scaling rules for industrial implementation.

Highly filled polyurethane (PU) formulations are widely used across industrial applications, yet their processing remains complex and difficult to scale. As a matter of fact, translating these materials from laboratory development to industrial-scale manufacturing remains challenging due to their complex rheology, high solids loading, elevated viscosity, and sensitivity to processing conditions. The scientific literature highlights that both powder blending and high viscosity blending suffer from the difficulty of even characterizing the material of interest, making fully predictive design and scale-up especially problematic [1]. In this study, we investigate how filler wetting dynamics, agglomerate breakup mechanisms, air entrainment and release behavior, and variations in impeller geometries collectively influence dispersion quality and processing time. Our findings demonstrate that no single traditional scale up criterion, such as constant tip speed, constant shear rate, or constant power per unit volume, adequately captures the behavior of the highly filled dispersions across different formulations. Instead, robust and transferable scale up requires an integrated framework that considers parameters such as filler type and concentration, impeller design, shaft positioning, and the evolving rheological profile throughout the mixing process. [1] Handbook of Industrial Mixing, intro, xliii

It is well known that Reynolds number differs between agitator tanks of varying sizes, even when using the same fluid. This is a significant challenge for engineers and researchers undertaking scale-up or scale-down operations. Although a crucial point, it is often subject to misunderstanding and incorrect handling/selection of impeller type and baffle etc. Under certain special conditions (laminar regime but low viscosity due to small tank) where viscous forces dominate, isolated mixing regions resembling Taylor vortices or Quet-type flows form in the gap between blade the wing and the wall [1]. This causes overall circulated flow is significantly implemented. Using MAXBLEND, the flow patterns occurring in low Reynolds number regime (Quette flow pattern, Taylor vortex pattern, spiral pattern, spiral plus separation pattern, etc.) have been organised and systematised based on Reynolds number and Taylor number. In particular, the conditions for the occurrence of poor mixing regions were clarified, along with the behaviour in the transition zone where good and poor mixing states coexist. It was demonstrated that flow patterns can be controlled.

This work investigates numerical modelling strategies to describe the hydrodynamics in a standard 1.5 mL Crystal16 vial, a small-scale crystallizer agitated by a magnetic stirrer. A fully three-dimensional (3D), transient model is first developed using the shear stress transport (SST) turbulence model. Model simplifications are then explored, including a 3D transient laminar formulation and a two-dimensional axisymmetric (2D) steady-state approximation. The 3D turbulent flow solution is validated against literature data obtained from a steady-state single reference frame simulation with a k-ε RNG turbulence model [1]; good agreement is obtained. Although the flow at 1000 rpm has previously been described as moderately turbulent, a 3D transient laminar simulation shows a similar vertical velocity field, indicating that turbulence modelling may not be required at all. A 2D axisymmetric, steady-state, turbulent model is then constructed using a momentum source approach [2,3], where impeller effects are represented as volumetric forcing terms. Although minor discrepancies appear in the velocity profile, this simplification remains justified due the substantial reduction in computational cost, from 30h (3D turbulent) to 7h (3D laminar) and 25s (2D axisymmetric turbulent), all on a standard laptop. This low-cost approach is attractive for studying crystallization in stirred vials, enabling simulation of hydrodynamical particle interactions and the development of population-balance crystallization models. [1] Achermann et al., Chem. Eng. Sci., 256, (2022) [2] Joshi et al., Can. J. Chem. Eng., 89, (2011) [3] Huang & Li, Nuclear Reactor Thermal Hydraulics and Other Applications (Chap. 5), (2013)

Transitioning from batch to continuous processing is a key strategy in process intensification. This study investigates gas–liquid slug flow in a high-precision horizontal Taylor–Couette reactor to enhance hydrodynamic stability and mixing performance. Experiments were conducted with varying gap widths and fluid viscosities. In low-viscosity systems, slug coalescence occurred at larger gaps, whereas high-viscosity fluids maintained stable slug structures. PIV measurements revealed that each liquid slug contained two to four steady Taylor vortex cells with asymmetric circulation. A calm inward flow region near the gas–liquid interface acted as a barrier to axial dispersion. Turbulent kinetic energy analysis showed localized peaks near vortex outflow boundaries, while fluctuations were strongly suppressed near interfaces. Unlike single-phase flow, which transitioned to wavy vortex flow at high rotation rates, slug flow maintained stable laminar Taylor vortex flow over an expanded operating range. These results demonstrate that gas slugs function as hydrodynamic barriers that suppress wave propagation, enabling intensified radial mixing while preserving axial confinement. The findings provide design insights for high-performance continuous reactors.

Gas dispersion into the liquid phase by using mechanically agitated reactors has been used widely in chemical, biochemical, and pharmaceutical industries. Accurate estimation of power consumption in the presence of gas as well as determination of gas-liquid mass transfer is essential parameter for reactor design. There are several investigations carried out with single radial turbine impeller and the correlation has been proposed [1]. Nonetheless, power dissipation and mass transfer have been studied with a limited range of configuration, and the correlation has been proposed based on the results from small laboratory to pilot scale vessel. In this study, the power consumption and volumetric gas transfer coefficient of two stages and three stages 6-blade Disk Turbines were examined from vessel size of 310 mm and 600 mm. The estimated correlation was proposed and compared to that data obtained from commercial scale vessels with inner diameter of 1500 mm.

Micro-fluidized beds (MFBs) are attractive intensified gas–solid contactors because of their high heat and mass transfer potential. However, strong wall confinement in small-diameter columns often promotes severe slugging, which deteriorates gas–solid mixing and limits stable operation. In this work, the mixing behavior of baffled micro-fluidized beds was investigated to evaluate the role of internal baffles in suppressing slugging and improving operability. A stereolithography-based 3D-printed reactor with an internal diameter of 4 mm and a height of 300 mm was studied in four configurations: an unbaffled column and columns containing 5, 10, and 30 internal baffles. Experiments were carried out with two Geldart B particles, Alpha-SiCB289 and OlivineB97.5, over a range of gas velocities. Mixing and flow dynamics were characterized using pressure fluctuation analysis, wavelet-based multiscale decomposition, time–frequency hotspot mapping, and high-speed imaging. The results show that internal baffling effectively fragments large slugs into smaller flow structures, suppresses mesoscale oscillations associated with unstable gas–solid mixing, and promotes a more homogeneous flow pattern. Flow regime analysis further indicates that baffling significantly broadens the bubbling and turbulent operating windows while narrowing the slugging regime. In addition, the minimum fluidization Reynolds number decreased by 23% for Alpha-SiCB289 and 37% for OlivineB97.5. These findings demonstrate that internal baffling is an effective strategy for improving gas–solid mixing in micro-fluidized beds and provide useful design guidance for intensified thermal and reaction processes.

Mass transfer plays a crucial role in fast chemical reactions. Poor mixing reduces conversion rates and product yields and negatively affects the distribution of reaction products when the reaction rate is fast relative to mixing times. Accordingly, fast test reactions are usually employed to characterise and quantify mixing, enabling assessment of the influence of flow regime on the operation of chemical reactors. Computational Fluid Dynamics (CFD) was used to simulate a competitive/consecutive reaction system, the Bourne test reaction, in a NETmix mesostructured reactor. A mesh independence test was performed to identify the largest element size that ensured that chemical reaction CFD results were independent of mesh resolution. The effects of Reynolds number, Re, and reactor topology on secondary product selectivity, quantified by a segregation index parameter, were investigated. Results show that increasing Re intensified molecular contact between reactants, hindering the formation of the secondary product. Three regimes were identified: a segregated regime, where no convective mixing in the NETmix chambers occurred; a chaotic regime, where advective mixing promoted the homogenisation of reactant species; and a turbulent regime, where product distribution was no longer micromixing-controlled. Topology studies revealed that NETmix configurations with spherical chambers and cylindrical channels exhibited less selectivity than those with cylindrical chambers and prismatic channels, achieving lower segregation indexes at similar Re values. Overall, the developed CFD model successfully describes the effects of reactor design and operating conditions on mixing-controlled fast reactions in NETmix reactors, providing insight into mixing phenomena from macro to molecular scale.

Rotary disc contactors (RDCs) are widely used for liquid–liquid extraction, but quantifying axial back-mixing in multi-stage units remains difficult. Here, CFD-generated tracer pulse responses are combined with a stage-wise N-CSTRs-in-series model with back-mixing to identify effective mixing parameters for a multi-stage RDC. Transient CFD simulations were performed for 25 operating points spanning multiple flow-rate ratios and rotor speeds in a five-stage geometry. Outlet tracer signals were converted to residence-time distributions (RTDs) via baseline correction, area normalization, and nondimensional time. RTDs were fitted using peak-weighted least squares with explicit tail handling to reduce bias toward long-time data. A local search over the back-mixing coefficient k, coupled with discrete exploration of the number of ideal stages N, yields robust fits and enables automated parameter selection across cases. The fitted curves reproduce both peak location and decay, and a weighted R-squared metric provides a consistent indicator of fit quality. Sensitivity analyses show that k largely controls peak shape, motivating localized k-segmented searches and fixed-N sweeps to stabilize estimation across operating conditions. For each condition, the workflow returns k, N, and mean residence time tau, summarized as heat maps to reveal operating-condition dependence. The framework is transparent, reproducible, and transferable to other staged liquid–liquid contactors, supporting model-based design and scale-up.

Hydrodynamic cavitation (HC) has evolved from a destructive phenomenon into a potent tool for process intensification, particularly for energy-efficient emulsification. While linear devices like venturi tubes are widely used, they require relatively higher pressure drop for cavitation inception and their operating life is often limited because of erosion induced by cavitation near solid walls. In contrast, the vortex-based cavitation devices (VD) utilize a tangential inlet to generate a strong vortex leading to early inception of cavitation and creating a cavitation zone away from solid walls. In this work we computationally investigate and elucidate the multiscale mechanisms driving oil droplet deformation and breakup in VD. This study employs a Large-Eddy Simulation (LES) approach coupled with a hybrid Volume of Fluid (VOF) and Discrete Phase Model (DPM) framework. This hybrid strategy combines VOF to resolve the transient deformation of the primary droplet interface on the finite volume grid, with DPM to track the dynamics of sub-grid liquid fragments post-breakup. The research investigates the influence of the Weber number on transient deformation topologies, specifically analyzing the transition from VOF-resolved ligaments to DPM-tracked daughter droplets. Results reveal that the complex interplay of anisotropic turbulence and pressure gradients within the vortex core drives critical deformation, leading to distinct sub-droplet size distributions. This comprehensive analysis provides new insights into the dynamics of emulsification under cavitation conditions. The presented approach, model and results provide a sound basis for understanding emulsification in VDs and thereby expand their applications to emulsification and other multiphase processes.

Traditional flow visualisation techniques such as Particle Image Velocimetry (PIV) and Positron Emission Particle Tracking (PEPT) are expensive and are essentially limited to specialist University Laboratories. Particle Tracking Velocimetry (PTV) is a 3D Lagrangian tracking technique for visualising fluid flow. Parallax setups limit the field of view in PTV systems. This study has addressed these issues by developing a low-cost optical PTV alternative based on a single-frequency light source and orthogonal mirrors to increase the field of view, which can be deployed in any open laboratory. The new PTV rig is presently optimised for a 300 mm stirred tank. Development was guided by finite element modelling of the system, including optical characteristics of the materials, enabling first-principles design and optimisation. This work discusses reconstruction-model development from first principles, approaches for handling large data throughput to obtain a more representative mixing study across a wide range of fluid mixing dynamics, and the use of optical mechanics to reduce signal-to-noise ratio. Conventional PTV methodology uses a narrow camera offset to minimise stereographic depth under obscuration constraints. As PTV does not have the same particle density as PIV, the camera offset need not be minimised; an orthogonal arrangement maximises spatial field while maintaining high spatial resolution. Finally, the benefits of combining the PTV digital twin with experimental PTV results in a stirred tank are discussed, with example results presented.

The monitoring of suspension, deposition, and re-suspension phenomena of aluminium-based adjuvants in pharmaceutical formulation tanks represents a significant technological and regulatory challenge, particularly when operating at low solid volume fractions. Aluminium phosphate adjuvants are widely used in vaccine formulations due to their immunostimulatory properties; however, they exhibit complex physicochemical behaviour in aqueous environments. These materials are electrically non-conductive and tend to form cohesive, gel-like sediment layers upon settling. Once deposited, such layers may undergo structural consolidation over time, leading to the formation of dense cakes whose mechanical strength and re-suspension behaviour depend strongly on formulation parameters. In particular, buffer composition, pH, ionic strength, protein adsorption onto particle surfaces, and storage duration have all been shown to significantly affect interparticle interactions and sediment strength. As a consequence, ensuring homogeneous suspension during manufacturing and prior to filling remains a critical quality attribute in pharmaceutical processing. Traditional methods for monitoring solid-liquid dynamics in stirred tanks, such as visual inspection, off-line sampling, or intrusive probes, are often inadequate for aluminium-based adjuvants. These approaches can be disruptive, lack spatial resolution, or fail to capture transient phenomena such as partial suspension or localized deposition near the vessel bottom. In this context, Electrical Resistance Tomography (ERT) offers a promising alternative as a non-intrusive Process Analytical Technology (PAT) capable of providing real-time, spatially resolved information on solid distribution and dynamics within opaque systems. ERT operates by injecting low-amplitude electrical currents through arrays of electrodes positioned on or within the vessel walls and measuring the resulting voltage differences. In solid-liquid systems containing non-conductive particles dispersed in a conductive liquid phase, local changes in solids concentration translate directly into measurable variations in electrical conductivity. In the present work, a bespoke ERT configuration was developed to enhance sensitivity to deposition and re-suspension phenomena occurring near the tank bottom, which are of primary relevance for pharmaceutical formulation processes. Experimental investigations were conducted at two different vessel scales to assess the robustness and scalability of the technique. A laboratory-scale vessel with a working volume of 7 L and a scale-down system of 300 mL were employed. Both vessels were equipped with a custom-designed linear electrode probe mounted at the tank bottom. The use of geometrically similar probes across scales allowed for direct comparison of ERT responses and facilitated the evaluation of scale-dependent effects. A systematic experimental campaign was performed to investigate aluminium phosphate settling, deposition, and re-suspension under a range of operating and formulation conditions. Impeller rotational speed was varied to explore different suspension regimes, from complete off-bottom suspension to partial suspension and full sedimentation. Deposition time was controlled to assess the evolution of sediment structure and packing density, while protein adsorption conditions were modified to mimic realistic pharmaceutical formulations in which antigens are adsorbed onto the adjuvant surface. The ERT measurements revealed distinct conductivity signatures associated with each regime, enabling clear discrimination between suspended, partially settled, and fully deposited states. The results demonstrate that ERT can reliably capture both the kinetics of aluminium phosphate settling and the progressive consolidation of the sediment layer over time. During re-suspension experiments, ERT was able to detect the onset of particle mobilization, the gradual erosion of the sediment bed, and the eventual restoration of a homogeneous suspension as impeller speed increased. Importantly, differences in re-suspension behaviour arising from protein adsorption and extended deposition times were clearly reflected in the conductivity maps, highlighting the sensitivity of the technique to subtle changes in particle-particle and particle-fluid interactions. Overall, this study confirms Electrical Resistance Tomography as a robust, non-intrusive PAT tool capable of providing both qualitative and quantitative insight into critical solid-liquid mixing phenomena under realistic pharmaceutical processing conditions. By enabling real-time monitoring of suspension quality, sediment formation, and re-suspension dynamics, ERT has the potential to support improved process understanding, enhanced control strategies, and more reliable scale-up of formulation operations involving aluminium-based adjuvants.

The understanding of multi-phase fluid dynamic phenomena in stirred tanks is essential for developing new processes as well as optimization and scale-up efforts. Therefore, these phenomena have been studied in the past using various methods, each with its own limitation: optical tomography is restricted to non-opaque systems, invasive probes provide only local information, and X-ray techniques require tracer particles to measure velocity. Magnetic resonance imaging (MRI) offers a powerful alternative, in which the magnetization of nuclear spins is measured to provide information on different phases and fluid velocity without the need of tracers. The MRI system at TUHH allows the operando investigation of process engineering vessels in relevant sizes up to 400mm in diameter with a modular stirrer setup, that can facilitate various stirred tank setups. In addition to studying gas-liquid systems, emulsions and suspensions have been investigated for the first time using MRI in this scale. Furthermore, local fluid velocities were directly measured using MR velocimetry at a spatial resolution of 2x2x2mm³. The phase distribution obtained by these measurements provide fresh insights into mixing and separation times, agglomerates formation and suspension distribution with sub-second temporal resolution. Three-dimensional flow maps and show the influence of stirrer types and speed as well as different mixture compositions. From these maps local shear stress information can be extracted, crucial to systems employing sensitive cells. These new insights allow the validation of simulation data and will help improving scale-up and scale-down models for stirred tanks.

Paste formulations, exhibiting non-Newtonian rheological characteristics such as shear-thinning behaviour, are a common format used to deliver personal care products. In a full formulation, high-shear mixing is used to achieve agglomerate break-up and powder dispersion into the continuous phase. It is of interest to study the effects of geometry, operating conditions and fluid properties on the flow field in such processes. Sawtooth impellers are a type of high-shear impeller used in paste manufacture. In this work, particle image velocimetry (PIV) was used to investigate custom sawtooth impeller geometries. Flow profiles were obtained for 3 impellers [4, 8 and 16 teeth] agitating various fluids [Newtonian – glycerol solutions (50 - 100 wt. %) and shear-thinning non-Newtonian – Carbopol solution (1.5 wt. %)]. Experiments were carried out at a rheometer scale [D = 15 mm and T = 35 mm]. Mean flow patterns in the vertical plane reveal the position of internal flow recirculation loops above and below the impeller. Deformation of the free surface is also observed as draw-down towards the impeller intensifies as agitation conditions change. Axial and radial velocity profiles are presented as a function of impeller surface area, viscosity and rotational speed. In the horizontal plane, flow patterns show fluid movement through the gaps between impeller teeth and the change in these patterns as the number of teeth increase. Understanding shear rates near and between impeller teeth will underpin studies into manufacturing industrially relevant formulations and their dependence on agitator geometry.

Micromixers are key components of microfluidics and are widely used in research and industry. Micromixers offer many benefits, such as rapid mixing, portability, safety, and low cost. Due to the low Reynolds numbers in microscale flows, the fluid is laminar, making it difficult to mix thoroughly, quickly, and effectively. In this study, 2D and 3D SAR-type micromixers composed of four hexagonal mixing units are proposed. Two variations of the Hexagonal mixer are designed, varying the aspect ratio (Α) and connecting angle (θ) in both 2D and 3D. The primary goal is to examine the effects of the aspect ratio and the connecting angle on mixing performance. The mixing index (MI), pressure drop, fluid flow, velocity, and mass transport are numerically analyzed employing Ansys Fluent 15 commercial software for liquid water at Reynolds numbers (Re) ranging from 1 to 200. It is evident from the numerical simulation that the mixing index depends significantly on both the aspect ratio (A) and the connecting angle (θ). In addition, the 3D Hexagonal mixer yields higher efficiency compared to the 2D mixer. The Aspect ratio equal to 1 and the connecting angle equal to 90 degrees provide the maximum efficiency, indicating 99% at Re = 200. Although the pressure drop of the 2D Hexagonal mixer is lower than that of the 3D one, the mixing energy cost (MEC) is much higher due to a lower mixing index (MI). Hence, the 3D Hexagonal mixer, especially with A=1 & θ=90°stood out as the best-performing mixer.

Mechanically agitated tanks are widely used in the chemical and biochemical industries. The flow developing in mixing tanks is highly complex and three-dimensional. Although few attempts have been made to locally measure the three components of velocity in the three spatial directions, the experimental techniques usually provide two-dimensional Eulerian data. Only recent developments (Shake-the-Box (STB) technique; Wieneke, 2013; Schanz et al., 2013, 2016) have made it possible to measure highly resolved spatial and temporal particle tracks for a high-density particle seeding (typically 0.05 particles per pixel). The objective of this study is to evaluate the accuracy of 4D-LPT in the turbulent regime and to understand its limitations through a comparative study with Large Eddy Simulation. In this context, 4D-LPT measurements are conducted. To the author knowledge this work is one of the first attempt where such a technique has been employed to study mixing tanks. It offers a unique opportunity to measure local and time-resolved velocity fields in a volume. The fluid investigated is water. The impeller is a classical Rushton turbine. Two mixing frequencies are analyzed (N∈[50,100], where N is the mixing frequency. Four cameras (Vision Research VEO 640) are used with 100mm Zeiss Milvius lenses with an aperture of F/16 and 540nm band-pass filters and Scheimpflug V3 mounts (LaVision). The investigated volume is homogeneously illuminated with a high frequency laser Photonics Industries DMX60-527-DH (2×60mJ at 1kHz). The measurement volume has the following dimensions: x=175.5mm, y=232.5mm, z=80mm, starting 40mm above the bottom of the mixing tank. First, Eulerian data are investigated. Local and time-resolved data are post-processed to evaluate ensemble and phase-averaged velocity fields in order to investigate organized structures. In particular, the development of coherent structures in the wake of the impeller is characterized. Furthermore, organized and turbulent motions are assessed through POD for both experimental and numerical data. Finally, Lagrangian data are studied with specific tools to 1) understand the flow dynamics at particle level and 2) characterize the loss of accuracy during the conversion from Lagrangian to Eulerian framework.

The mixing characteristics within a stirred tank are influenced not only by the geometry of the mixing system but also by the impeller Reynolds number (Re). In particular, it is well known that poorly mixed regions can form depending on the value of Re. Mixing time serves as an indicator of the time required for the entire tank, including these poorly mixed zones, to reach homogeneity. Correlation equations based on this definition have been widely reported. However, in processes where rapid mixing of added substances is required — especially when uniformity in the well-mixed region is critical — there is a need for correlation equations that specifically represent mixing time in these well-mixed regions. Despite this practical need, such correlations have not yet been reported. In this study, the concentration gradient (∇C), defined as the spatial gradient (∇) of concentration (C), was introduced as a key parameter to characterize mixing. Based on this, mixing time was redefined as the time required for ∇C to become uniform throughout the stirred tank. Using this definition, a correlation equation was derived from the convection-diffusion equation. The validity of the proposed correlation was evaluated using reported data across laminar, transitional, and turbulent flow regimes. The results indicate that the derived equation effectively predicts mixing time under a range of flow conditions.

In high-viscosity mixing, deformation accompanied by stretching/compression and repeated reorientation plays an important role; however, quantitative methods for evaluating mixing performance have not yet been fully established. This study proposes a performance-evaluation framework for distributive and dispersive mixing based on both numerical analysis and experiments by combining a macroscopic mixing index with a local deformation-mode metric. The target system is a twin-screw mixing device in which the rotation direction and rotational speed of each screw can be independently controlled. The flow patterns are broadly classified into two operating modes: co-rotating conveying and counter-rotating conveying. For flow analysis, the Moving Particle Fully-Implicit (MPFI) method is employed to accurately evaluate the coupled behavior of rotation and deformation. In the simulations, mixing indices based on the redistribution of initially labeled particles are evaluated using either region-based definitions or neighborhood statistics within a prescribed length scale. In addition, local deformation modes, such as shear-dominant and extension-dominant behaviors, and their temporal histories are analyzed using metrics derived from deformation and rotation information. These analyses clarify the mixing characteristics of each operating mode. In the experiments, some Non-Newtonian fluids are used. The time evolution of dye concentration and the number fraction of mixed iron powder are measured. By relating these results to the simulated mixing indices, we discuss mode-dependent differences in mixing performance and their correlation with deformation-related metrics.

Lamellar gel networks (LGN) are multiphase structures found in a range of personal care products including hair conditioners. Evolving fluid rheology and changing fill level during manufacture present significant mixing challenges in production and in scale-down for discovery of new formulations. This work explores the potential of a high throughput testing device for the investigation of new formulations, employing a helical ribbon in a 34 mm diameter cup filled to a height of 36 mm (100%). Since flow characteristics at this scale are not normally representative of those at pilot and production scale, this work aims at linking the scales through characterisation of the fluid dynamics and mixing. Dynamic torque measurements employing Newtonian and shear-thinning model fluids has shown that power draw in the laminar regime is generally linear with respect to rotational speed at fill levels of 25, 62.5 and 100%. In the laminar regime, inverse proportionality of the power number to the Reynolds number is found (3 < Re < 24), as expected. Comparison with OpenFOAM CFD simulations using the SIMPLE algorithm at low rotational speeds, where the free surface is modelled as flat, are in good agreement. However, as the rotational speed increases, vortexing and deformation of the free surface occurs in the experiments leading to an overprediction of torque, and consequently power draw. Volume of fluid (VOF) simulations allowing for free surface deformation give improved predictions, further validated by 2-D Particle Image Velocimetry (PIV) data obtained in both horizontal and vertical planes.

High quality of the mixing plays a decisive role in rapid reactions, affecting the quality of products. For comparison, the quality of micromixing and the specific energy dissipation rate in microreactors for following geometry and methods of supply were investigated: feeding through two tangential pipes (TU+TU), through two radial pipes (RU+RU), through a tangential and axial pipe (TU+C), and through a radial and axial pipe (RU+C). Furthermore, a special case of feeding solutions simultaneously through three pipes was investigated: two tangential and an axial pipe (TU+TU+C), as well as through two radial and an axial pipe (RU+RU+C). This case corresponds to feeding three solutions into the first chamber, which is used in some types of synthesis of nanosized particles of inorganic materials. Based on the research results, the following conclusions were made: 1) The use of a swirling (feeding into tangential pipes) allows one to achieve higher quality of micromixing than in the absence of a swirling (supply to radial pipes) for all methods of feeding; 2) The primary contact of the reagents in the zone of the neck results in higher quality of micromixing than that for the primary contact of the reagents in the upper chamber zone (i.e. the micromixing in TU+C and TU+TU+C is better than TU+TU, RU+C and RU+RU+C is better than RU+RU method); 3) At high flow rates, the quality of micromixing for the supply in three inlet pipes (TU+TU+C method) MRISF-2 is better than that for the TU+C method.

High pressure homogenization is widely applied in the food industry to disrupt cell structures in plant based beverages. Achieving sufficient disruption to prevent grainy mouthfeel often requires energy inputs that may be unsustainably high. This limitation is linked to an incomplete fundamental understanding of cell structure breakup, partly due to previous studies overlooking the heterogeneity of plant materials and the differing susceptibility of distinct cell types to homogenization. In the present study, plant based cell suspensions were homogenized across a broad range of conditions, including variations in raw materials, pressures, and number of passages. Samples were analysed using automated microscopy, and particle images were classified through a convolutional neural network (CNN) trained to distinguish morphological categories. This approach enabled quantitative assessment of how individual cell types respond to homogenization. The classifier demonstrated high precision and reliably differentiated particle morphologies. Electron microscopy provided additional confirmation by linking morphological categories to specific cell types. Results revealed clear differences in homogenization susceptibility between cell types. When combined with an analytical framework previously used for inorganic particles, the findings indicate distinct breakup mechanisms, including rupture for aleurone cells and erosion for pericarp cells. These insights contribute to a deeper understanding of cell structure disruption and offer guidance for optimizing the design and operation of high pressure homogenization processes.

As for liquid-liquid two-phase flow with mass transfer in macroscopic extraction equipment, droplet size distribution (DSD) is the key factor for transfer efficiency. Droplet behaviors, affecting the DSD, are correlated with interfacial tension which varies with the mass transfer process. To understand the evolution of the DSD under mass transfer, the droplet breakage kernels and dynamic interfacial tension (DIFT) were investigated. First, to obtain and use reliable breakage experimental data, the breakage kernel measurement methods in literature were summarized and reexamined. It was found the measurement definitions of breakage frequency in the most studies are not correct. Based on the correct combination between breakage fraction and total spent time, the suitable measurement standards for single droplet and droplet swarm experiments were established. Second, through analyzing droplet deformation and breakage process, the droplet collides with multiple eddies during its breakage time. The interaction between eddies and the droplet causes the droplet to oscillate, and the second-order oscillation breakage mechanism was determined. Reexamining three classic breakage constraints and introducing the concept of breakage similarity, the corresponding breakage model was constructed with experimental data, and the model exhibits satisfying prediction for droplet breakage in macroscopic extraction equipment with turbulent flow. Third, based on the relationship between breakage frequency and interfacial tension, a DIFT measurement method in a stirred tank was established by conducting droplet swarm experiment with a high-speed camera. The effects of impeller rotating speed, solute concentration and residence time on the DIFT were investigated. Besides, the DIFT is mainly determined by interfacial solute concentration and the method predicting DIFT was constructed with mass transfer results obtained by analyzing single droplet dispersion process. Finally, some studies concerning computational fluid dynamics - population balance model (CFD-PBM) were conducted with the correlations of breakage kernels and DIFT.

Drop breakup is a vital process of producing emulsions across industries and flow regimes. The effect of turbulence on drop breakup has been extensively studied, likewise, how emulsifiers lower the interfacial tension between phases, affecting the coalescence and initial breakup in laminar flows. However, the combination of the two is less explored - how emulsifiers affect the drop breakup in turbulent flows where the drop breakage is much faster than diffusional transport. Some suggested mechanisms, beyond lowering the interfacial tension, show that under some conditions the emulsifier makes the drops more difficult to break. The mechanisms have been described as an increase in effective viscosity [1] or interfacial tension [2]. These kinds of elastic effects have also been seen in a model of a homogeniser, a highly turbulent emulsification device. In this study, we investigated the effect of emulsifiers on drop breakup by screening different emulsifiers with emulsifications experiments using a batch rotor-stator mixer and a low disperse phase volume fraction oil-in-water emulsion with different concentrations of the emulsifier. The results show that there is an effect of the emulsifier that cannot solely be explained by interfacial tension as there is a stabilising elastic effect at intermediate concentrations in some conditions. We also report on how the effect depend on the molecular properties of the emulsifier. References [1] Padron & Calabrese, 2023. Chem. Eng. Res. Des. 195, 447. [2] Janssen et al., 1994. Colloids Surf. A. 91, 141. With support from the Swedish Research Council (VR 2024-04823).

Crystallisation is widely used in the fine chemical industries to produce solid products with desired properties. Crystalline products exhibit a particle size and shape distribution, which affects downstream processing (e.g., filtration) and product quality. Designing reproducible crystallisation processes that deliver a target PSSD is therefore critical. In practice, non-uniform mixing and spatial heterogeneities in hydrodynamic properties, such as energy dissipation rate, can influence crystallisation. While coupled CFD–population balance models have been developed to explore these effects computationally, systematically derived experimental datasets for quantifying the effect in reality and model validation remain limited. This study examines mixing effects in seeded batch cooling crystallisation through experimental investigation. Unlike conventional studies that focus primarily on particle size distributions (PSD), this work quantifies the evolution of particle size and shape distributions (PSSD) under varying hydrodynamic conditions. Experiments are performed in stirred tank crystallisers of different scale and geometry, with impeller speed and suspension density systematically varied. PSSDs are measured using imaging-based characterization devices and population dynamics are probed at different locations through focused beam reflectance measurement (FBRM). Spatial variations in solute concentration are quantified through local sampling, using ATR-FTIR spectroscopy, and gravimetry. Additionally, local suspension density, impeller-shaft torque (for power input), and slurry viscosity are measured. Reproducibility and scalability are assessed by comparing crystallisation behaviour and product properties across scales using defined scale-up criteria. The experimental insights provide a quantitative basis for understanding how hydrodynamics govern crystallisation outcomes and aid the digital design of reproducible and scalable crystallisation processes.

Applications of CFD to stirred tanks in the laminar and turbulent regimes are ubiquitous, yet many tanks used for the manufacture of formulated liquid consumer products of complex and/or evolving rheology operate in the transitional regime, where “turbulence structures” are difficult to capture. Recent work has demonstrated good agreement between experimental 2D-PIV data and CFD simulations for Newtonian fluids using a Stress Blended Eddy Simulation (SBES) closure, obtained for a stirred tank equipped with a A320 hydrofoil impeller, over a wide range of Reynolds numbers in the transitional flow regime. SBES is a hybrid model developed by Ansys employing a RANS k–ω SST model close to the wall boundaries and LES in the bulk. Previous in-silico studies on non-Newtonian fluids in the transitional regime have used RANS or LES approaches but not SBES. In this study the use of SBES (Ansys 2025R2) is extended to non-Newtonian fluids in the transitional flow regime. The simulations were carried out in a tank diameter, T = H = 0.19 m, the impeller diameter, and clearance being D = T/2 and C = T/3, respectively. The fluid used was a shear-thinning 0.6% wt aqueous carboxymethyl cellulose solution that gives a Reynolds number of approximately 1900. Flow fields obtained with SBES and a RANS model (GEKO) were compared with 2D-PIV data, with the SBES model showing the best agreement.

Mixing profiles of formulated liquid and solid products such as soap is complicated by an evolving non-Newtonian rheology during saponification, which creates considerable difficulty for optimising process conditions and energy use. Poor mixing can result in an incomplete saponification process and a poor-quality end product. Whilst current work by Hart et al has led to the development of a digital optimisation workflow using simplified rheology, scope remains for improving the rheological model to more accurately represent processed material and final product. The main challenge in measuring soap rheology is the presence of slip on the rheometer geometry walls, making collection of data difficult with poor repeatability and reproducibility. Being unable to measure the rheological profile of soap limits the industry’s ability to optimise process, equipment design and process time which collectively limit the ability to control process energy input. This paper describes a scoping exercise to screen the capability of different instruments to measure soap rheology, to thus establish a validated rheological model for the digital workflow. This rheological model will then be compared with torque measurements obtained from a 5-blade ploughshare mixer, modelled on industrial mixers. This mixer contains regions of high shear, low shear, radial flow, and axial flow in close proximity, representing a more complex flow field. This study will thus determine the capability of the torque measurements to determine progress towards process endpoint, using the improved rheological model as a baseline to see whether in-situ determination of the product rheology matches expectations

Dissolution-based chemical recycling can recover high-quality polymers, but scaling remains challenging. Under specific operating conditions, polymer dissolution can be described as externally mass-transfer-controlled, neglecting internal solvent diffusion and polymer swelling (1). For instance, this study (2) focuses on mixing miscible systems with strong physical properties differences, using concentrated glucose syrup as model fluid dissolving in water. This system behaves analogously to a polymer-solvent and hydrodynamic mixing regimes are identified, highlighting limitations of classical correlations and extending applicability. This study investigates how hydrodynamics governs polymer dissolution. Rheological measurements, including frequency and time sweeps within the linear viscoelastic regime, assess whether polypropylene dissolution is controlled by internal polymer dynamics or external mass transfer. Measurements are performed at temperatures around 160 °C and polymer mass fractions between 40 and 70 wt%, representatives of industrial conditions. Assuming rheological equivalence between non-homogeneous systems and homogeneous mixtures at the same average polymer concentration, an effective diffusion coefficient is identified and used to construct Deborah and Sherwood numbers. In this way, rheological insight provides the basis for subsequent hydrodynamic analysis, with relevant mixing conditions explored through dimensionless groups such as Re* and Ri* and mass-transfer coefficients estimated to support dissolution optimization and scale-up. References 1) Martini, R. E., Brignole, E. A., & Barbosa, S. E. (2009). Dissolution mechanism of polymers in high pressure–high temperature n‐alkanes—Application to blends separation. Polymer Engineering & Science, 49(3), 602-612. 2) Mirfasihi et al., Chemical Engineering Journal 486, 149712 (2024).

For the first time, experimental and numerical simulation was conducted on the mass‑transfer and key hydrodynamic characteristics of a novel type of impellers – centrifugal‑pulsed impellers (CPI) in a single and double configurations within a stirred‑tank reactor (STR). CPI generate pulsations due to the specific geometry of their surface, transforming rotational forces into local pressure pulsations. These pulses form a dynamic velocity field dependent on both amplitude and frequency, which are governed by the design features of the impeller’s curved‑shape surface and its rotational speed. This leads to reduced energy dispersion through a decrease in large‑scale turbulence within the vessel, an increase in droplet and bubble dispersion extension, improved mixing quality within the dispersed phase (especially inside droplets), and increased kinetic coefficients (i.e., heat and mass transfer). These improvements are achieved by means of localized, predominantly near‑impeller, directed energy input and reduced axial load on the impeller shaft. The pulsation mechanism was broken down through mass‑transfer coefficients and hydrophone power spectral density decomposition into key blade‑passing frequencies and their harmonics. The power number was experimentally determined, and twenty‑six configurations of impeller modifications were examined through computational modelling. Mass‑transfer coefficients, velocity fields, and specific energy‑dissipation profiles were evaluated for each CPI design, including blades with sin‑phased and counter‑phased double‑impeller configurations. It has been revealed that orifices in the impellers significantly increase local turbulence in the regions below and above the impellers by generating additional directed flows, leading to enhanced circulation throughout the entire STR volume.

Tubular photobioreactors for microalgae cultivation suffer from strong radial gradients in light intensity, mass transfer limitations, and progressive biofilm formation on transparent walls, all of which reduce productivity and operability. This contribution investigates the use of static and dynamic mixers inserted into tubular photobioreactors to tailor hydrodynamic conditions with three primary objectives: (i) homogenization of the light field within the illuminated liquid layer, (ii) intensification of mixing and mass transfer for efficient utilization of nutrients and CO2, and (iii) enhancement of wall shear stresses to mitigate biofilm growth on transparent tube surfaces. A set of mixers with variable geometry was designed, enabling systematic variation of key geometric parameters relevant for flow pattern, residence time distribution, and wall shear. The performance of mixers was evaluated through a combination of numerical and experimental methods. In order to characterize flow structures and hydrodynamic indicators relevant to light and mass transfer, numerical simulations were employed. Furthermore, laboratory-scale experiments were conducted in order to quantify mixing/homogenization efficiency and pressure drop across the mixers. The resulting data were then compared with those obtained from a commercially available mixer. The present study explores the trade-offs between mixing intensity and hydraulic losses, and identifies geometry ranges where enhanced homogenization and increased wall shear can be achieved without prohibitive pressure effects. The findings provide a rational basis for the design and scale-up of tubular photobioreactors equipped with static or dynamic mixers, with the aim of achieving more robust and energy-efficient microalgae cultivation.

In this contribution, bionics, understood as the connection between biology and technology, is applied in the context of an unconventional feasibility study on the design of novel agitators. Bionics attempts to transfer phenomena from nature to technical applications. Therefore, first, a proven and sensible biological principle must be found and analysed, and second, a technical application must be identified in which the biological principle found can be usefully applied. In order to find a solution to a flow problem, an approach from nature that has proven successful over millions of years but has not yet been used for mixing in technical systems is to be employed here. The aim of this work is to find out whether the underlying natural phenomenon is also applicable to stirring. Both a whale fin-like radial and an axial stirrer are being developed and characterised. The newly designed stirrers are manufactured on a laboratory scale using a 3D printer and tested in several series of experiments. These tests are always carried out in comparison with conventional stirrers (pitched blade and Rushton turbine), on the basis of whose dimensions they were designed. The exploratory experiments carried out involve measuring the power characteristics and performing exemplary mixing tasks such as suspension, homogenisation and emulsification. In the course of the study, the stirrers are further optimised iteratively based on the findings obtained from the experiments. The study presents the results of this unconventional, sustainable development process and aims to encourage creative, new, efficiency-oriented stirrer designs.

The Taylor-Couette (TC) reactor represents a distinct type of stirred tank, where mixing is driven by the relative rotation of concentric cylinders. The cylinders generate vortical flows that govern mixing, yet the mechanistic role of gap width (the annular space between concentric cylinders) remains under explored. This study investigates how gap width (radius ratio η = Rᵢ/Rₒ) regulates mixing time and mechanisms via tracer experiments and CFD simulations. Results reveal that gap width critically modulates Taylor vortex structure and mixing efficiency. Narrow gaps (η → 1) delay Taylor vortex onset, requiring higher Taylor numbers for mixing initiation; once formed, vortices are axially compressed, enhancing intra-vortex circulation but limiting radial exchange. Wide gaps (η → 0) promote early vortex formation with larger, axially stretched vortices that enhance axial dispersion but weaken local shear and energy dissipation. Mixing time exhibits a non-monotonic dependence on gap width: optimal efficiency occurs at intermediate η (~0.6-0.8), where intra-vortex circulation and inter-vortex exchange reach a balance. Below this range, excessive axial dispersion prolongs homogenization; above it, confined vortices restrict global mixing. Turbulent kinetic energy dissipation rate profiles confirm that maximum energy efficiency aligns with this optimal window. These findings establish gap width as a key design parameter for TC reactors, providing a mechanistic framework for tailoring mixing performance in applications such as fermentation and nanoparticle synthesis.

In the cooling crystallization process for potassium chloride production from seawater-derived concentrated brine, small crystal growth rate results in small crystals and poor solid–liquid separation efficiency. To address this issue, a draft tube (DT) is often placed in stirred vessels to control the vertical flow and promote particle agglomeration. Herein, the impact of the DT geometry on the solid–liquid hydrodynamics was investigated to identify the mechanisms promoting agglomeration and determine the optimal scale-up criteria. Single-phase flow was simulated using computational fluid dynamics (CFD) with the SST k-ω turbulence model. The simulations were validated by comparing the velocity profiles with particle image velocimetry measurements. The spatial crystal size distribution and crystal concentration in multiphase flow were evaluated using an Eulerian-Eulerian approach incorporating a granular model coupled with a discrete population balance model (PBM). The top-mounted impeller generated an upward flow inside the DT, enhancing particle suspension within the DT region. Compared with a straight DT, an inverted conical DT with a steeper gradient (top–bottom = Φ60–30) promoted particle suspension and internal recirculation. These flow characteristics are expected to increase the frequency of gentle particle–particle collisions, thereby promoting agglomeration. Furthermore, the upward flow rate was optimized by adjusting the rotational speed, leading to increased floating from the bottom of the vessel and greater crystal suspension inside the DT. Experimental crystallization using the Φ60–30 DT also demonstrated superior performance relative to the straight DT, indicating that the proposed simulation is a reliable tool for crystallizer design.

Higher hydrocarbons are considered highly promising for storing surplus electrical energy produced from renewable sources. Fischer–Tropsch synthesis (FTS) is a well known process for the catalytic conversion of synthesis gas into higher hydrocarbons, which can subsequently be upgraded into sulfur free, aromatics free fuels and chemicals. For low temperature FT synthesis, slurry bubble column reactors and multitubular fixed bed reactors are typically used. Owing to their many advantages (phase mixing, small temperature gradients, excellent heat management, and equipment without moving parts), slurry bubble column reactors are employed in industrial plants (e.g., by Sasol and Exxon). For industrial use of bubble column reactors, challenges arise in scaling up these reactors and in separating solid catalysts from liquid products. For these reasons, this type of reactor continues to be investigated. The aim of this contribution is to present a comparison of predictions from various computational correlations for gas–liquid–solid three phase systems published in the literature for slurry bubble column reactors, using a model system corresponding to an industrial FTS process.

Heterogeneously catalyzed reactions employing solid catalysts or enzymes are often carried out in suspension reactors. This reactor type has two main disadvantages. First, catalyst separation is time‑ and cost‑intensive and is associated with a loss of catalyst. Furthermore, enzyme‑catalyzed reactions are limited by the rotational speed to avoid mechanical damage to the enzymes by strong shearing. Hollow‑shaft stirrers are conventionally used for gassing processes. For the first time, the present work combines the characteristics of conventional hollow‑shaft stirrers with the principle of heterogeneous catalysis to eliminate the aforementioned problems in terms of sustainability and cost optimization. For this purpose, the hollow shafts of two newly designed stirrers are packed with catalyst, and instead of gas, the reaction medium is passed through the catalyst-packed hollow shaft. In a first step, the characterization of the designed stirrers is performed using two complementary approaches: (1) Model reaction with ion exchanger in the stirrer shaft and a solution of sodium chloride as reaction medium and (2) power characteristics to evaluate the power input and determine the optimum rotational speed. To investigate the effectiveness of the new designs, comparative experiments are carried out in a suspension reactor using two traditional stirrers. The results show promising performance of the hollow-shaft stirrers to catalyst retention, while maintaining or even reducing energy consumption. The resulting elimination of catalyst separation leads to substantial savings in processing time and operation costs and ensures more sustainable industrial mixing processes (by minimizing catalyst loss).

Continuous-flow mixers operating under chaotic advection enable efficient mixing of both highly viscous and low-viscosity liquids, including those containing components sensitive to high shear rates typical of turbulent flow. Chaotic advection ensures rapid development of intermaterial surface area and a reduction in segregation scales, thereby accelerating mass transfer across the intermaterial surface. However, numerical modelling of mixing in such systems using computational fluid dynamics (CFD) requires very fine computational meshes to accurately resolve steep concentration gradients in regions where fast chemical reactions occur. When the local segregation scale becomes comparable to or smaller than the computational cell size, numerical diffusion smears concentration gradients and significantly affects the predicted reaction outcome. In this work, a hybrid mixing model was applied to simulate mixing in a tubular reactor operating under chaotic flow conditions. Premixing in the vicinity of the injection point for the limiting reactant was modelled using CFD. In the downstream region, where chaotic advection dominates, mixing was modelled in the Lagrangian framework by solving the reactants mass balance equations for deforming liquid striations. The segregation scale and reagent concentrations obtained in the initial CFD-based stage were used as initial conditions for the Lagrangian calculations. By comparing modelling results with experimentally determined selectivities of competitive-parallel reactions, the average rate of deformation of fluid elements in chaotic flow was determined as a function of the Reynolds number.

Filamentous fungi are shear-sensitive microorganisms. Depending on the growth and flow conditions, they either grow in submerged cultures as loose mycelium or form agglomerates known as pellets. The production of natural substances can be supported depending on the growth form, and the growth form also has an important influence on the process itself. Loose mycelium increases the viscosity of the broth, which can lead to poor mixing and, consequently, oxygen and nutrient limitations. It can also lead to overgrowth of probes and other components in the bioreactor. Cultivation in pellet form does not present these problems; however, due to the dense network inside the pellets, limitations in mass transport can arise, leading to an undersupply of the core. In the case of Penicillium sp. (IBWF 040-09), the pellet form has proven suitable for producing a protease inhibitor, since this secondary metabolite is produced when the supply of substrate is limited. To specifically influence the morphology, different types of stirrer were used during cultivation, and the resulting morphology was examined. A Rushton turbine, a 3-segment blade stirrer and a modified 3-segment blade stirrer with a ring were used for this purpose. CFD simulations will be carried out in cooperation with another working group to determine the shear forces and these will be validated by mixing time experiments. This will create a simulation-based basis for further optimization of the stirrer geometry. The morphological changes caused by the different agitators have already been demonstrated and are currently being investigated with regard to important properties such as the circularity and density of the hyphal network inside the pellets.

The hydrodynamics of Gas-Liquid columns significantly influences both the intensity of mass transfer (kLa) and energy consumption, expressed by specific energy dissipation rate. The balance between these two factors defines the optimal geometry of inserts and optimal regimes. Hydrodynamics of the novel column equipped with swirlers working in parallel and generating unidirectional vortices was studied experimentally and theoretically. From one to four swirlers have been located in various positions of the supporting tray with variations of distance between them, and eight configurations of swirlers’ positions have been compared. The main idea of this work was to study a novel column equipped with swirlers working in parallel and generating unidirectional vortices, intensively interacting one with the other, thus leading to increased turbulent shear stresses and to further improved bubble disintegration and better mass transfer rate. A comparative analysis of a column with parallel swirlers and traditional gas-liquid equipment used in chemical, petrochemical, and biotechnology industries was performed. The maximum value of εav is approximately 70 times higher than that for stirred-tank reactors and up to 100 times higher than that in bubble columns. The obtained data allowed to predict increased gas disintegration efficiency and a higher mass transfer coefficient in Gas-Liquid systems. In our further experimental studies, the influence of liquid and gas flow rates on Gas-Liquid mass transfer characteristics and specific surface area will be performed.

Microorganisms are used to produce a wide range of compounds, including biopolymers. For this purpose, both wild-type and genetically modified strains can be utilized. One example is genetically modified E. coli, a strain used for the production of P3HB, a polymer belonging to the PHA group. The source of organic carbon in such bioprocesses is often glucose; however, it may also consist of wastes derived from industry or agriculture. The aim of this study was to investigate the effects of mixing on the optimization of the growth of genetically modified E. coli used for the production of poly(3-hydroxybutyrate) under aerobic conditions, with glucose as the primary carbon source. An additional objective was to select an appropriate (radio)tracer for determining RTD. Such investigations are particularly important due to the possibility of bacterial growth at high cell density, where significant oxygen limitation may occur. The measurements were carried out in two bioreactors, BioFlo 310 and 415, with aeration and working volumes of 5 and 10 dm³, to study the scale effect. The influence of impeller speed and air flow rate on the oxygen transfer was analyzed. The measured kLa values were compared with correlations reported in the literature, and modifications to the empirical expressions were proposed to achieve better agreement with experimental data. The results obtained may be used particularly for: 1) validation of CFD models; 2) determination of the enhancement coefficients under aerobic microbial cultivation; 3) intensification of microbial growth and their metabolites, including the potential for scale-up.

This work focuses on a novel method that provides a unique solution to the closure problem occurring in reactive turbulent flows. So far, closure models have been developed for both single and competitive reactions, such as competitive–parallel and competitive–consecutive schemes; however, more complex systems have not yet been properly analyzed. In particular, this work focuses on the diazo-coupling test reaction, which is one of the most commonly recommended chemical probe approaches for investigating micromixing. The closure is based on an assumed mixture fraction function for a non-reactive scalar, accurately describing mixing of fully segregated streams. The method was further validated with carefully selected mechanistic micromixing models, and experimentally confirmed in studies using an in-line rotor–stator high-shear homogenizer. The results clearly demonstrate that the model accurately describes the course of the diazo-coupling reaction; it enables the correct determination of the contribution of by-products in the reactive mixture, which serves as a key indicator of local micromixing kinetics. The method introduced in this work, combined with CFD, will enable, in the near future, a comprehensive assessment of the integrated effects of convection and mixing at the macro-, meso-, and microscale in industrially relevant equipment.

Solid–liquid separation is a fundamental unit operation in chemical and process industries. Conventional filtration methods often suffer from clogging and maintenance issues, particularly when handling fine particles, limiting their applicability in continuous processes. This study proposes a novel hydrodynamics-based separation concept utilizing the unique flow characteristics of a Parallel Paddle® (PP) impeller. The PP impeller consists of two parallel plates and generates a tornado-like swirling upward flow beneath the impeller even at relatively low rotational speeds. When a perforated partition plate with two openings was installed in the upper region of the stirred tank, an unexpected particle segregation phenomenon was observed. Fine calcium carbonate particles accumulated above the partition plate, while the lower region of the vessel became significantly clarified. To understand this behavior, a process engineering model was developed based on particle mass balance and flow field structure. The separation mechanism is interpreted as the interplay between upward swirling flow, localized circulation zones, and gravitational settling. A simplified compartment model was constructed to describe particle transfer between upper and lower regions, enabling quantitative prediction of separation performance. The results demonstrate the feasibility of a non-filter-based fine particle separation device driven purely by controlled hydrodynamics. This approach offers potential advantages for continuous solid–liquid processing without mechanical filtration components.

This work adopts the Lattice Boltzmann Method (LBM) coupled with Large Eddy Simulation (LES) to conduct high-fidelity simulation of turbulent mixing and liquid-liquid dispersion dynamics in Rushton baffled stirred tanks, combining the Eulerian-Lagrangian approach and Discrete Phase Model (DPM) for droplet tracking. A D3Q19 multiple-relaxation-time LBM scheme is employed to solve the continuous phase flow field, with the immersed boundary method for accurate modeling of complex tank geometries including rotating impellers and baffles. To address the key challenge of Lagrangian time step dependence in traditional droplet breakage models, a probabilistic breakage approach is proposed, where breakage probability is proportional to the time step and breakage occurs when the probability exceeds a random 0–1 value. This model eliminates artificial breakage frequency distortion and accurately predicts the Sauter mean diameter and droplet size distribution. LBM simulations reveal detailed mixing characteristics: droplet trajectory tortuosity increases with impeller speed due to enhanced radial/axial circulations, peak droplet velocity locates in the impeller zone and shifts toward the tank wall at higher rotational speeds. The LBM framework enables precise capture of multi-scale mixing hydrodynamics, including turbulent energy dissipation and droplet motion, providing a reliable numerical tool for understanding and optimizing stirred tank mixing processes in chemical and process engineering.

Batch rotor-stators are used in several dispersion processes that require the breakup of immiscible droplets, particle clusters or for foam generation on account of the high turbulence energy dissipation rate in the vicinity of the mixer head. The overall process performance also relies on minimal spatial variation of product properties. Blending performance of rotor-stators, on the other hand, has not been widely studied/reported. In this study, the blending performance of three different mixer head geometries of the VMI Turbotest batch rotor-stator has been determined evaluating mixing time values on the basis of power input and compared with predictions from the turbulent mixing time correlation. Turbulent power numbers were determined as 3.36, 2.47 and 2.62 for the slotted wide gap, gridded and slotted with toothed rotor mixer heads respectively. Analysis of mixing time results has shown that the dependence of mixing time on power input in the turbulent blending regime holds for all three mixer heads, as with impeller driven set-ups. A brief study with a pitched blade turbine gave results in agreement with predictions from the mixing time correlation. With the VMI Turbotester the correlation also held in its general form but with a slightly higher constant of 5.9 instead of 5.2. It could be concluded that in the turbulent regime these batch rotor-stators can achieve sufficient bulk blending, i.e. can be used for dispersion processes without the need for an additional impeller for bulk circulation.

This study investigates the mixing performance of a newly designed impeller developed to improve homogenization efficiency in high-viscosity systems while maintaining the grinding and dispersion performance characteristic of the Cowles-type geometry. Experimental observations indicate that the limited homogenization performance of the conventional Cowles impeller in viscous media primarily originates from geometric constraints that restrict the generation of strong bulk circulation and axial flow components. Based on these findings, a novel impeller geometry was designed and manufactured to enhance macro-mixing while preserving the high shear regions required for grinding and particle dispersion. The performance of the new impeller was evaluated through comparative mixing trials. Homogenization efficiency was assessed using colorimetric analysis and visual uniformity of mixtures produced under comparable operating conditions. The results demonstrate that the newly developed impeller significantly improves homogenization in high-viscosity systems while maintaining comparable grinding and blending performance to the conventional Cowles impeller. These findings suggest that appropriate modifications to impeller geometry can substantially enhance mixing efficiency in viscous formulations without compromising dispersion capability. The impeller geometry investigated in this study is protected under the European patent titled “Mixing Tank Impeller,” granted by the European Patent Office under patent number EP4454741 on 10 September 2025. The patent is assigned to Kansai Altan Boya Sanayi ve Ticaret A.Ş., and the author of this work is the inventor of the patented impeller design. The experimental study presented in this paper reflects the scientific evaluation and laboratory-scale validation of this patented mixing technology.

n many industrial syntheses — particularly in pharmaceutical manufacturing and the production of fine chemicals — the characteristic timescales of the reactions are comparable to, or even shorter than, those of the mixing process. When multiple reactions proceed simultaneously, the local mixing behavior can therefore exert a significant influence on reaction selectivity. The reliable use of such reactions requires a dedicated reactor design. First, extremely short mixing times — particularly micromixing times — are essential to ensure a rapid incorporation of the reactants into the bulk fluid at the point of feed inflow. Achieving such mixing times requires high local energy dissipation rates. Second, a high degree of process control is necessary to maintain reproducible mixing behavior and, consequently, consistent selectivity. Since in the pharmaceutical industry multipurpose plants are commonly used to produce various products in batch mode, additional process flexibility is required. This includes the ability to handle different process conditions such as fluctuating viscosities, varying flow rates, and multiphase systems within the same equipment. Over the past decades, static mixers have become well-established as reactors offering excellent micromixing performance and a high degree of process control. However, due to their inherent design, they provide only limited flexibility with respect to varying process conditions. For this reason, stirred tank reactors continue to be the predominant choice in industrial practice. Yet, their comparatively long mixing times make them only partially suitable for mixing-sensitive reactions. Currently, alternative reactor types that combine the flexibility of stirred tanks with short mixing times characteristic for static mixers are lacking in the pharmaceutical industry. In this work, two innovative reactor concepts for mixing-sensitive reactions are investigated, each aiming to unite the flexibility of stirred tank reactors with short mixing times: 1. the Reaction Mixing Pump (RMP) 2. the Jet Loop Reactor (JLR) To investigate the meso- and micromixing behavior of these reactors, the influence of various process parameters such as feed rates and feed positions on the selectivity of the established Villermaux–Dushman test system was examined under single-phase operating conditions. Based on the experimental results, micromixing times were calculated using the engulfment micromixing model. In addition, this model was used to predict the expected selectivities of this synthesis in the reactors. The experimental results demonstrate that even small changes in process parameters can have a significant impact on reaction selectivity. For both reactors, a reduction in feed rate led to a transition from the mesomixing to the micromixing regime. For example, in the case of the RMP, very low feed rates resulted in an increase of the mixing time, indicating backflow into the feed capillary. Micromixing times in both reactors were determined to lie in the range of 10⁻⁴ to 10⁻² seconds. Using the micromixing model, the reaction selectivity in the micromixing regime could be predicted reliably. The results demonstrate that both the RMP and the JLR can achieve micromixing times comparable to those of specialized static mixers, while offering the process flexibility of stirred tanks. Overall, these findings indicate that the RMP and JLR are highly promising reactor concepts for achieving high selectivities in mixing-sensitive reactions under variable process conditions.