ARR - Annual Report

Present annual report-2002 described results on the second year’s work after renewal of the first period (1998-2000) in the theme of “Mechanochemistry of Materials” approved by the IFPRI organization. The work focuses on structural change of cellulose by grinding and its dissolution in aqueous NaOH solution. Under this concept, the contractor (F. Saito) and his group have extensively investigated the mechanochemical (MC) work from both viewpoints of experimental and computer chemistry. Several characterizations such X-ray diffraction (XRD), TG-DTA and FT-IR analyses have been conducted for the ground sample, and regarding the XRD analysis, the grinding enables us to transform the structure of cellulose into disordered system like an amorphous state in the prolonged grinding. The TG-DTA and FT-IR spectra are almost the same as those in the initial stage, the dissolution of cellulose in the aqueous NaOH solution is however improved. Cellulose has donor and accepter composing hydrogen bonds in the molecule, and the bond is strong relatively in comparison with other similar structured materials such as glucose and lactose. In spite of this, improvement in the dissolution of cellulose in the solution may be due to the disturbance of internal molecules of hydrogen bonds and their rupturing. This is followed by the investigation on the computer chemistry work, which has shown the weakest part of the chain structure of cellulose in the all, so that the rupture may be initiated to break at C-O bonds in the molecule. Of course, other parts of the structure are also subjected to damage by the grinding, however, the unit cell like cellubiose may remain unaltered in the prolonged grinding. This may be confirmed from the results of TG-DTA and FT-IR analyses. The substance re-crystallized from the solution may be cellulose, of which structure may have slightly different from the initial one. Grinding the cellulose results in the agglomeration of particles, of which sizes are not significantly changed. These facts suggest the suitable application of the ground cellulose to food and pharmaceutical fields.

This work has been presented at the IFPRI AGM 2002 held at Sendai, Japan in 15-18, July 2002. At this moment, it is still unknown, but “mechanochemistry of cellulose” has a potential to extend to food and pharmaceutical fields, because the ground sample of cellulose is, in fact, still maintaining the same unit cell even in the prolonged grinding stage, changing its part of the structure, correspondingly its physico-chemical properties such as solubility. During the course of this investigation, the evidence may be initiation to start an application of mechanochemistry to food and drugs field by modification of morphology and structure of such products. This is also a great expectation of this series of work to come.

According to the project work objectives there were the following four aims to be achieved in the second year of the project extension:

1. Carrying out validation of drying tests in the counter-current spray drying tower
2. Carrying out experimental small-scale drying kinetics tests on selected products
3. Comprehensive validation of the CFD model for scaling-up spray drying process in co- and counter-current system
4. Carrying out experimental investigations to identify the effect of initial process parameters (feed properties, feed rate, air flow rate, drying temperature, atomization parameters) on the final product properties (porosity, bulk density, solubility, etc.)

This project has the goal of providing experimental evidence for the influence of interparticle surface forces and hydrodynamic forces on the moderate to high shear rheological properties and shear stability of wet dispersions that span the colloidal to particulate range. The current emphasis is on particle shape and its effect on the shear thickening transition, as well as the dynamics of the shear thickening transition. Ongoing research into the influence of polymer stabilization, and in particular on the role of adsorbed polymer in modifying the hydrodynamic and interparticle forces acting between particles under flow will not be summarized here, but in the next year’s report.

PART I

In PART I of this report, the transient shear rheology (i.e. frequency and strain dependence) is compared to the steady rheology for a model colloidal dispersion through the shear thickening transition. Reversible shear thickening is observed and the transition stress compares well to theoretical predictions. Steady and transient shear thickening are observed to occur at the same value of the average stress. The critical strain for shear thickening is found to depend inversely on the frequency at fixed applied stress for low frequencies (high strains), but limits to an apparent minimum critical strain at higher frequencies. This minimum critical strain is shown to be an artifact of slip. Lissajous plots illustrate the transition in material properties through the shear thickening transition, and the energy dissipated by a shear thickening suspension is analyzed as a function of strain amplitude. This work provides crucial evidence for the dynamics of the shear thickening transition, and also suggests that dynamic oscillatory measurements can be used to study shear thickening in addition to steady shear.

PART II

PART II provides experimental evidence for the effects of varying particle shape on the shear rheology and shear thickening in concentrated colloidal dispersions. A series of 2:1, 5:1 and 8:1 aspect CaCO3 dispersions were prepared, characterized, and examined rheologically. Prelimary SANS measurements corroborates the supposition that alignment can reduce the low shear viscosity and severity of shear thinning at higher particle loadings. In comparison to previous work on charged silica dispersions, these dispersions shear thicken at lower stresses, but still shear thicken at higher stress than would be expected for hard sphere dispersions at the same loading and average particle size. Ongoing experimentation and modeling is designed to fully elucidate the influence of particle shape on the low and high shear suspension rheology.

This report summarises progress in IFPRI project 37 in 2001/2. The breakage of wet granules in a granulator environment is being studied by Rachel Smith (UQ PhD student). Preliminary results from the dynamic compression of single wet pellets using an Instron Dynamite load frame are reported. The methodology developed by Iveson and Page is used with sone improvements. Fast frame video is used to capture in detail the deformation of the pellets.

The deformation behaviour of the pellets varied widely from very plastic behaviour to failure by brittle crack propagation. Brittle or semi-brittle behaviour was more likely at high strain rate and with high viscosity binders. Pellets made from non-spherical, broad size distribution lactose powder were also more likely to fail in a brittle mode than closely sized glass ballotini pellets. Preliminary results suggest the failure mode can be related to a critical capillary number. The fast frame video analysis of the compression is a very useful tool for understanding the failure behaviour of these partially saturated materials. A hypothesis for wet granule breakage in mixer granulators is suggested and a plan for testing the hypothesis using a specially designed breakage only granulator is presented.

A summary of the PhD studies of Hans Wildeboer on regime separated granulation is given. This work proposes that to optimise granulator design, the key rate processes for granulation should be separated. A conceptual design for a two stage granulator – nucleation only, followed by consolidation and layered growth - is presented. Mathematical models for both stages are developed. The models are compared to results from a partly regime separated continuous drum granulator. The models are very promising in predicting and understanding the granulation behaviour.

A new granulator design is developed which completely separates the nucleation regime from consolidation and growth. A novel nucleation device is tested which gives near mono-sized nuclei at a required size. Further details of Wildeboer’s work are given in his PhD thesis, which will be sent to IFPRI members in early 2003.

The process of disintegration of liquid/solid suspension jets and sheets by atomization is analysed in a fundamental manner and visualized by suitable measurement methods, which allow qualitative and quantitative evaluation of the process. Supporting numerical analysis and theoretical derivations will contribute to basic understanding and control of the suspension atomization process. Model suspensions will be atomized by means of conventional and specifically designed atomizers. The fourth year activities that are reported here include:

• Experimental investigations of suspension atomization in twin-fluid atomizer
• Experimental investigations of suspension atomization in rotary-atomizer

Model suspensions based on water with industrial relevant suspended particles (China Clay) have been atomized by means of twin-fluid atomizer. Model suspensions based on water and water/CMC-(carboxymetylcellulose) mixture with suspended glass particles have been atomized by means of a rotary atomizer.

EXECUTIVE SUMMARY

Our basic concept in the project was as follows: for a fundamental discussion on the phenomena of electrostatic charging of powder, it is essential to study the charge generation due to a single impact/contact on a single particle. To realize the concept we proposed two directions:

Approach 1: Impact Charging Experiment

To extend the original (previous version of the) impact charging experiments to the actual powder particle size region, a new equipment was designed and built in the previous project years. The sensitivity of the charge measurement was extremely improved, and this allowed particles from 50 to 500 μm to be applied to the experiments. Although the actual data were dispersed, almost all of them fell in a certain region explained by the charge relaxation model, which determines the amount of impact charge non-empirically, with an assumption of localized initial charge. Particles which were plated with nickel to make its surface electro conductive lost their all the initial charge at the impact onto a metal target as the same to the cases of 3 mm metal particles. Experiments with different kind of metal targets did not show any significant difference due to the difference of work function of the metals whilst it is expected that the impact charging behavior may depend on it if the process of charge relaxation due to gas discharge is not considered. These facts encouraged us to say that the charge relaxation model works in this particle size region of hundred micro-meter as well as the case of bigger particle with 1-3 mm in diameter in the previous work. In the future study, more improvement of the sensitivity of the charge measurement will be tried to allow particles under 10 μm to be applied. As well, an experimental study chasing process of successive impacts will be interesting.

Approach 2: Development of the Method of Electrostatic Adhesive Force Measurement

This is the completely new approach to determine the amount of charge transferred onto a particle due to a contact from a measurement of adhesive force curve at approaching and separating particle against a metal target. We have actually launched AFM study: force curves of particles due to electrostatic attraction were successfully observed with PS particles of 30 and 100 μm, and with glass bead with 40μm. In the cases of smaller particle, the force curves on approach and separation were almost same after enough repeated contact, whilst they were not in the case of bigger particle: the 100 μm particle case showed that the force in the case of separation was significantly bigger than the case of approach even after enough repeated contacts. This proved that some kind of charge relaxation takes place: this must be investigated in more detail in the future study. To evaluate the force curve quantitatively, a disk charge interaction gave a good account, and it was indicated that the amount of charge and the contact area can be estimated by the method.

The fluidised dense-phase (FDP) conveying of powders and low-velocity slug-flow (LVSF) of granular bulk solids are the most common and popular modes of dense-phase used in industry. However, the accurate prediction of conveying performance still is not possible from first principles and relies heavily on empiricism.

The main aim of this project is to develop the necessary understanding, databases, guidelines and models for the purpose of predicting accurate optimal operating conditions for the two modes of dense-phase. However, as was mentioned in the original research grant application, it was unlikely that both the FDP and LVSF sections could be completed thoroughly in a single 3-year period (i.e. due to the amount of work involved). Hence, top priority was given initially to the LVSF section of the project, although some progress was also made with the FDP section of work. However, with the 3-year extension to the research grant a substantial amount of work can now be completed in the FDP section, as well as tying off some loose ends from the LVSF section.

Several difficulties have been encountered during the course of the project (e.g. unexpected results and phenomena) and have delayed progress in various areas. In some cases, it was not possible to complete certain scheduled tasks (e.g. testing aluminium and mild steel pipe and wide range of granular solids). In other cases, it was necessary to pursue new work (e.g. rotary valve air leakage, new pipe friction and stress transmission testers). However, in terms of achieving the main goals, there is no doubt that the project will be successful in terms of improved understanding and the development of new databases and models for the prediction of LVSF performance. Unfortunately, due to the various problems and delays to date, the full range of pipe wall materials and bulk solids will not be able to be tested – such work is necessary to confirm the accuracy and validity of the new models (e.g. majority of work to date has concentrated on poly pellets). Also, a significant amount of additional time will be needed for the relatively more complex FDP section of work (e.g. only one product and a few different pipelines were able to be tested by the end of the initial 3-year period). The 3-year extension is allowing a more concentrated effort in this area, as evidenced by the extensive testing completed over the past 12 months.

This Annual Report summarises the research progress and major achievements to date, as well the forward plan for the next 12 months.

Based on the Fokker-Planck-Equation a new model was developed to calculate the reduction of periodic concentration fluctuations entering a continuous mixer. The calculations and the experimental results show:

1. The average residence time normalized with the period length of the entering mass flow fluctuations is the main influencing parameter of the mixing quality.
2. The mixing process can be described very well with a system of Fokker-Planck-Equations.

By the development of an air-cleanable cowl for the Near Infrared (NIR) Spectrometer VECTOR 22/N it is possible to determine in-line the mixing quality at the end of the continuous mixer every 1.7 sec. The NIR Spectrometer can be used for the analysis of any powder mixture, the only requirement is that the tracer component contains C-H-, O-H- or N-H-groups which absorb the NIR radiation. Because the analysis results can only be as good as the calibration results the calibration was optimised and an apparatus for dust free, easier and faster calibration of the NIR Spectrometer was developed.

For calcium carbonate with an average diameter of 2 μm it could have been shown that the pulsation of a volumetric feeder with a single proportioning device can be reduced tremendously by the use of different dosing tube attachments. The best results were obtained with a self-developed rotating star attachment. The pulsations were reduced to a tenth of the pulsations of a standard dosing tube.

The increasing inter-particle interactions with increasing fineness are a serious problem during wet grinding in stirred media mills. Because these inter-particle interactions have a big influence on the stability against agglomeration and on the rheology of the suspension. As shown in the literature [1] and in this report, a lower boundary of possible particle size is typically reached at around 1 μm. Furthermore the agglomerate size can even increase with further grinding, building larger particles than the original ones, that are strong enough to withstand the comminution process.

In this work we postulate that colloidal stability must be considered in wet grind- ing to understand there results and surmount there limitations on the production of nano- particles.

To study the grinding limits of particle sizes below 1 μm a detailed understanding of the agglomeration process and its mechanism is needed. Therefore the agglomeration process of fine particles is discussed in this report first under Brownian and later on under turbulent shear motion similar to the motion in stirred media mills. The agglomeration kinetics as function of added salt was measured using dynamic light scattering (DLS). Further information on the agglomeration process and the structure of the agglomerates give small angle neutron scattering (SANS) experiments.

Theoretical predictions on the stability ratio and the critical coagulation con- centration based on the DLVO-theory were calculated using measured values of the ζ-potential.

Furthermore an initial investigation to describe the equilibrium between break- age, agglomeration and deagglomeration in stirred media mills has been conducted. The problem is described by the population balance model, which will be solved using the moment method.