ARR - Annual Report

The investigations of the last five years have shown that a production of stable suspensions of hard ceramic materials with a median particle size smaller than 10 nm is possible by wet grinding in a stirred media mill. Nevertheless, no absolute limit of the grindability was yet reached.

By means of experimental results the influence of the electrostatic stabilization on different pH-values during the comminution of fused corundum on the grinding progress and the grinding media wear is discussed.

Results for comminution of fused corundum with grinding media of different diameters between 100 μm and 1300 μm and stirrer tip speeds between 6 m/s and 15 m/s are presented and discussed with regard to the grinding progress and the grinding media wear.

The particles in stirred media mills are stressed and ground between two grinding beads, one grinding bead and a wall or between a grinding bead and an agitator element in zones of high energy density [1, 2, 3]. Only for a desagglomeration or a desintegration of microorganisms shear forces are a further grinding mechanism.

During the approaching process of two grinding beads or a grinding bead and a wall the fluid between the collision partners will flow out of the gap. The smaller the product particles are, the better they will follow the fluid out of the active volume being the volume in which particles can be captured and stressed [4, 5].

With decreasing product particle sizes two opposite effects can be observed:

• While two grinding media approach each other, the fluid between them will be displaced. Wereby particles will follow the fluid flow out of the gap more easily the smaller the particles are. This may lead to media contacts without any grinding.

• On the other hand, with decreasing product particle size the absolute number of particles increases with the reduction factor to the power of three. Thus, a capture and stressing of more than one particle or a particle bed is possible.

The loss of kinetic energy during the collision is partly used for comminution. It has to be asked, how the energy is transferred to the feed particles. For that reason it is interesting to know how many particles are captured. According to the number of captured particles three cases can be distinguished [6]:

a) Only one particle is captured, which is stressed with the total energy (single particle stressing).

b) More than one particle is captured between two beads, all particles have contact to both beads during the stress event and all particles are stressed independent of each other. In this case at first that particle is captured, which has the largest size and/or which has the smallest distance to the connection line of both bead centers. This particle is stressed with the maximum energy. The particles, which are captured between the two beads after the first particle, are stressed with a considerably reduced energy. At the end of the stress event diverse single particle stressings with different intensities occur.

c) A particle bed is captured and stressed between two grinding beads.

Up to now it is not investigated how many particles are captured and stressed.

For a better understanding of the stress events in the sub micron particle size range, the particle motion in the displacement flow between a grinding chamber wall and an approaching grinding media was observed. Due to the small sizes of grinding beads and especially of product particles the product particle motion cannot be investigated optically in existing (real sized) stirred media mills. For that reason a model was developed. Parameters as volume concentration, approaching velocity and angle, model grinding bead diameter and material were varied.

Nanotechnology applications in the pharmaceutical, materials, and chemical industries has renewed interest in the use of wet grinding in stirred media mills for the production of nanoparticles. However, challenges arise in the production of sub-micron particles that are, in part, due to colloidal surface forces influencing slurry stability and rheology. As often observed in the literature, a grinding limit in the range of 0.5 μm is reached despite high energy inputs and aggressive milling conditions. Furthermore, the product size can even increase with increased energy input, a seemingly counterintuitive result that may be attributed to aggregation of fine particles during the comminution process. In this work we postulate that colloidal stability and rheology must be considered in wet grinding to understand these results and to surmount limitations on the production of nano-sized particles. Experiments are performed on a well-characterized, model system of monodisperse primary nanoparticles that are destabilized and aggregated under various milling conditions. Conditions spanning Brownian to turbulent collision aggregation in a model stirred media mill are explored to study the effects of colloidal stability on the aggregation process.

The agglomeration kinetics are measured using dynamic light scattering (DLS) as a function of particle and electrolyte concentrations. Further information on the agglomeration process and the structure of the agglomerates are also obtained from small angle neutron scattering (SANS) experiments both at rest and under flow. Theoretical predictions based on independently measured particle and solution properties as well as mill characteristics are compared against the experimental results to demonstrate that particle aggregation kinetics in a stirred media mill can be controlled by tailoring colloidal interactions and the milling conditions. Furthermore it is shown that the concept of electrostatic stabilization during wet grinding of nanoparticles can also be applied to the system of tin oxide. It is shown that in contrast to alumina no mechanochemical changes occur for the system of tin oxide during the wet grinding process. Thus the obtained median particle sizes are the result of pure mechanical grinding. In addition the suspension rheology in a stirred media mill as function of grinding time and inter particle interactions is studied.

This research provides a theoretical basis for understanding stirred media milling of nanoparticle slurries and as such, is a step towards a predictive model of nanogrinding in stirred media milling.

Summary of 3rd year

In this part of the project work a new mixing/extrusion concept was further developed to study the microstructure changes of a model system, which undergoes a transition from a 3- phase wet powder to a 2-phase concentrated suspension system (3P2S Transition) (Arancio and Windhab, 2003).

The further developed model system used consists of:

• powder/binder: liquid fat melt/oil (e.g. cocoa butter)(i)
• powder/solid filler: hydrophobic silica particles (Sip.D17, Degussa, (D)) (ii)

The first advantage of using a fat melt, which is in the fully liquid state at moderately elevated temperatures (eg. 40°C for cocoa butter), was given by the ease of solidification close to room temperature. This made investigations of the wet powder/suspension microstructure possible without additional structural changes during sampling from the process and during sample preparation. The second advantage proved was the good homogeneous mixing ability of the hydrophobic silica powder particles within the fat binder, even at extremely low binder fraction due to the fact that the binder fat melt was cold sprayed, thus producing small solidified binder particles. These solidified binder particles were mixed with the filler particles under NIR-based homogeneity control. If this powder mixture was heated up to 40°C, a significantly better homogeneity of the mixture of powder and liquid binder was obtained compared to the conventional/alternative process of mixing by spraying the low fraction of liquid binder into the filler powder (formation of local lumps).

The performed process consisted of the following steps:

1. spray cooling of the binder-fluid phase in order to obtain a solid fine binder powder
2. mixing of this binder powder with the filler powder and mixing quality control by near infrared spectroscopy
3. pre-compaction of the model system “powder/binder powder mixture (PBPM)” in a mechanical testing machine (Zwick)
4. ram extrusion of the pre-compacted system
5. micro-structural investigation of the extruded product

The powder/binder mixture in the solid powder state of the binder (A) was used to investigate the local stresses/stress distribution leading to deformation of the solid cocoa butter (binder) particles. This allows identifying the zones where the transformation from a three phase wet powder to a two phase concentrated suspension (3P2S) preferably happens. The solid liquid mixture above the melting temperature of the binder (cocoa butter) (B) reflects the powder binder system in the normally applied state, with air additionally entrapped. The acting pressure and shear stresses lead to a synergistic reduction/compression of the air included until the 3P2S transformation takes place. The impact of pressure, shear and pressure + shear on the resulting powder/binder microstructure were investigated in detail and 3P2S phase diagrams derived, which form a new basis for design calculations of extruders/extrusion processes and extrudate property optimization.

The aim of this project is to enhance our understanding of how solvents affect the polymorphic forms of compounds. Polymorphism is very important to the pharmaceutical industry as a compound’s polymorph can exhibit different physical properties, some more desirable than others. Being able to control which polymorphic form is nucleated is thus very useful.

To study the effects of solvent type on polymorphism, enantiotropic compounds will be used. These compounds have multiple polymorphs but the distinguishing feature about them is that there is a transition temperature where the solubilities of the polymorphs are equal. This transition temperature is constant regardless of the solvent type as at this temperature the free energies of the polymorphs are equal. This allows us to study the effects of solvent type by decoupling it from the effects of the supersaturation.

The work done in this first year is primarily aimed at identifying the compounds which are enantiotropic in nature, selecting a model compound, characterizing it and determining its exact transition temperature. Subsequently, we will perform experiments using different solvents at the transition temperature and measure both induction times and nucleation rate.

Project Overview

In the first year of this project, a multidimensional population balance model has been developed for a general granulation process. The model details the evolution of particles with respect to distributions in size, porosity, and binder (moisture) content the three critical properties that are related to processing as well as end product properties.

Model Utilization

The model has been utilized to determine the control-relevant properties of the granulation process. The influence of changes in the binder addition rate as well as the agitator speed were determined using sensitivity analysis of the population balance model.

Results and Insights

The results provide insight into the degree to which size, porosity, and moisture content can be adjusted in a feedback control model of operation. The present analysis is flexible enough to address both batch and continuous granulation.

Implications

The insights generated from this analysis point to more effective strategies for the improved operation of industrial granulators, notably with regard to reduced recycle streams (both ¯nes and oversized particles).

The aim of the project is to establish a relationship between the product properties and feed material and the mill functions for milling of organic solids. The specific objectives are:

1. to characterise the physical, mechanical, and thermal properties of model organic (feed) materials (material function) at the single particle level, and to examine the effects of temperature and humidity on these properties
2. to investigate the breakage behaviour of single organic particles under various conditions
3. to characterise the properties of milled products, and to correlate the product properties to material and mill functions Models materials that are planned and approved for use in the project by the TC of IFPRI include aspirin (low hardness, relatively high fracture toughness, ductile), á-lactose monohydrate (áLM, high hardness, high fracture toughness, semi-brittle), sucrose or sorbitol (high hardness, relatively low fracture toughness, regarded as brittle in the literature), starch (ductile and mechanical properties greatly affected by the strain rate), and microcrystalline cellulose (MCC, medium hardness, high fracture toughness, semi-brittle). These materials cover a fairly wide range of physical, mechanical and thermal properties, hence ensuring generality of the results to be achieved. This report summarises the work done over the first year on five model organic materials, á-lactose monohydrate (áLM), sucrose, sorbitol, starch and microcrystalline cellulose (MCC). The work includes extensive experimental investigation into the behaviour of both the single particle impact breakage and the bulking milling under various conditions, and preliminary mathematical modelling based on the distinct element method (DEM) and the population balance method. Also included in the report are some results of the milling of MCC and áLM obtained by the investigators before the project was started. The single particle impact experiments provide data of the breakage extent of single particles as a function of impact velocity, which is then used to infer the physical and mechanical properties of the tested particles by utilising the Ghadiri-Zhang model for semi-brittle materials. The results show that the modes of failure of the five tested materials agree well with that reported in the literature. Sucrose and sorbitol, which are generally regarded as brittle materials, demonstrate the highest extent of breakage. On ii the other hand, starch and MCC, two materials regarded as more difficult to mill, show the lowest extent of breakage. Sub-ambient impact experiments are conducted on áLM, sorbitol and MCC. The results suggest that the breakage propensity, thus the mechanical properties, of áLM and MCC are little affected by the temperature, but a lower temperature is seen to reduce the breakage extent of sorbitol particles. The bulk milling experiments are conducted in a single ball mill and the results are quantified by an analogy to the first-order rate process. The results show that milling of both MCC and áLM is little affected by the temperature, in agreement with the single particle testing results.
4. Attempts are made to relate the single particle impact breakage behaviour to the bulk milling behaviour. For MCC, starch and áLM, the milling behaviour expressed by the milling rate constant relates linearly to the single particle parameter containing the physical and mechanical properties of the particles. However, the relationships for sucrose and sorbitol are non-linear, indicating alternative methods should be explored for these types of materials. The relationships between the single particle breakage and the bulking milling behaviour are different for different materials. Preliminary efforts are made to unifying these relationships. This is achieved by quantifying the mill function with a term called ‘milling power’ predicted by DEM simulations of motions of both the milling media and the feed particles. The new bulk milling parameter is shown to relate to the single particle behaviour well and a unified relationship is obtained for áLM, starch and MCC. Preliminary modelling work using the population balance method has been carried out on the milling of sucrose in collaboration with Dr R Bertrum Diemer of DuPont. The results will be reported in the future.

Nanotechnology applications in the pharmaceutical, materials and chemical industries has renewed interest in the use of wet grinding in stirred media mills for the production of nanoparticles. However, challenges arise in the production of sub-micron particles that are, in part, due to colloidal surface forces influencing slurry stability and rheology. As often observed in the literature, a grinding limit in the range of 0.5 µm is reached despite high energy inputs and aggressive milling conditions. Furthermore, the product size can even increase with increased energy input, a seemingly counterintuitive result that is attributed to aggregation of fine particles during the milling process. By producing particles smaller than a median particle size of 1 µm a steady state between breakage and agglomeration exists in the milling process. This equilibrium is controlled by interparticle interactions as well as the milling conditions. In this report we demonstrate that this steady state particle size is independent of the feed particle size and can be reached by agglomeration of small particles as well as by real breakage of large particles.

Furthermore, we extended our studies to non-aqueous systems, because of the high industrial demand on nano particles in organic solvents. Milling studies with steric and electrostatic stabilization showed that the size of the particles milled in organic solvents is conspicuously smaller than the size of the particles milled under the same conditions in the aqueous phase. However, ions dissolved from the media wear and impurities in the suspension can destabilize the suspensions if the electro-static stabilization mechanism is chosen. This is not the case for steric stabilized systems.

An important question is whether the mechano-chemical activation which was observed in aqueous media is crucial to obtain nano-particles. Differential Scanning Calorimetry (DSC) and X-ray difraction (XRD) measurements showed no mechano-chemical changes during milling in ethanol and toluene. In the aqueous phase the stabilization mechanism has no influence on the amount of hydroxide phase. A protecting polymer cover around the particles does not prohibit the mechano-chemical changes.

The present research is oriented towards the grand challenge to understand general powder dynamics and the ultimate goal of the work is to develop a quantitative description of active flows of fine powders. The study is centered on the “intermediate” regime of flow where both frictional and inertial effects are important and where fluctuations of strain rate and stress are significant. The main application is in the area of small-size, rough and/or cohesive powders that are industrially relevant. The novelty of the project is to study a relatively large range of materials and flow geometries to gain meaningful insight as opposed to limiting the work to a single “relevant” flow device using a “standard” powder. Extensive previous work mainly in the Physics literature on smooth glass beads and hard, metal spheres have done little to shed light on the behavior of industrially important powders that are usually non-spherical, rough, fine, cohesive and compressible.

We report on a series of materials from simple (round beds) to complex (fine, odd-shaped and slightly cohesive), used in a set of judiciously chosen equipment with interchangeable boundary conditions and measure stresses and stress fluctuations as a function of geometry and shear rate. The goal is to develop constitutive equations for powder flows and to use them in continuum-type theoretical models to predict flow patterns, velocity distributions and forces on boundaries such as stationary walls and moving pedals.

The approach follows the earlier work of Tardos, (1997), Tardos et al., (2003), and Tardos and Mort, (2005) and applies it to the geometry of the Couette device and to more complex flows such as hopper (funnel) and centripetal (“spheronizer” and high-shear mixer) flows that are relevant to storing, feeding, mixing and granulating powders.

This project has an overall goal of designing model-based control algorithms using multi-scale descriptions of particle attributes in high-medium shear granulators. The first objective of the second year of this project was to design a reduced order model that is appropriate for real time applications. The second objective involved a feasibility study of applying a reduced order model in a Model Predictive Control (MPC) algorithm. The methods used to reach these objectives have prepared us for the next stage, the objective for the third year: application of MPC algorithms on a pilot plant granulator. During the scope of this project four different granulation models have been considered (in order of complexity):

1. Transfer function model (Pottmann, et al., 2000; Gatzke and Doyle, 2001)
2. One dimensional well mixed discretized Population Balance Model (PBM) (Sanders, et al., 2005)
3. One dimensional three compartment discretized PBM (Wang, et al., 2006)
4. Three dimensional PBM (Immanuel and Doyle, 2005)

The models were either used as a plant model to simulate a real plant or as the basis for a linear controller model. This report described the details of the research with the second model. The fourth model has been described in our first year report (IFPRI # ARR51-01). The third model has been developed by the group in Queensland and had been adjusted to model their pilot plant, which will be used in future studies in this project. The main differences between model 2 and model 4 are the recycle flow (+ screens and crusher) and the introduction of three separate regimes in the granulator. More details about the UQ model and pilot plant are described in section 6: Summary and Future Work.

Abstract

The first part of this report (Chapter 3 and 4) present the framework that was used to setup and test an MPC controller on a simulated plant model. The plant model is based on earlier work done at The University of Sheffield and Organon, The Netherlands (Sanders, et al., 2005). The granulation kinetics were measured in a 10 liter batch granulator with an experimental design that included four process variables. The aggregation rates were extracted with a Discretized Population Balance model. Knowledge of the kinetics was used to model a continuous (well mixed) granulator. The controller model is a linearized state space model, derived from the nonlinear DPB model. It has the four process variables from the experimental design and a feed ratio as input variables. Since the DPB model describes the whole size distribution (GSD), different sets of output variables were chosen and compared. Preliminary results show successful implementation of Model Predictive Control on a continuous granulator. Further research is needed to test the linear model on an actual pilot plant.

Funding for this project has only commenced 10 days ago, as such no results are presented in this report. A brief report is presented of the preparations that have taken place. This primarily involves the acquisition of appropriate materials for the study.