FRR - Final Report

Executive Summary

Project ARR-98 focused on mixing rules for powder mixing in rotating drum flows. The first phase of the project explored the viability of PEPT measurements to yield the desired measurements for validating and complimenting a granular flow model of rotating drum flows. In this regard, coarse graining strategies proved integral to the success of the project. A useful outcome was the observation of the reverse Brazil nut effect observed for a binary mixture (by size).

Phase two built upon the serendipitous finding of a Peclet-based rheology that underpinned most of the granular flow phase space. ~200 DEM simulations across a wide range of flow configurations was used to successfully validate the new rheology.

For this purpose, the tetrahedron approach by Löwer et al. was applied, which provides a good approximation of the distribution of the capillary pressure as a function of the degree of saturation to laboratory values. Without knowledge of the capillary pressure curve of the particle system, the capillary pressure range between a fully saturated and a fully de-watered bulk structure can thus be estimated over a wide range of contact angles < 90°, but also > 90°.

The variation of wettability served as a parameter study. For this purpose, hydrophobically coated Al2O3 (aluminum oxide) particles are suspended in different liquid solutions. By changing the ethanol content of the otherwise aqueous suspension, the wetting behavior could be specifically adjusted. The influence of changes in the local contact angle within the cake structure could be demonstrated for many of the particle and filter cake properties investigated, which resulted in a model approach to reproduce the dewatering properties.

The measurement of dewatering characteristics involved analyses of filter cakes built up in the in situ cell followed by dewatering and subsequent measurement in μCT. After the measurement, a morphological analysis of the filter cake structure includes the local particle geometry as well as the inverse pore geometry. The analysis of the partially dewatered filter cakes allowed the measurement of the local contact angle distribution along the wetting line of the hydraulically isolated liquid regions within the pore space.

X-ray computed tomography allows non-destructive investigation of the process-relevant structural effects of the filter cakes, such as obtaining local information from within the filter cake, i.e., spatially resolved data instead of integral parameters.

The wettability of the solid surface within the filter cake structure significantly determines its dewatering properties. Associated with the wettability, several properties of the filter cake change, such as its pore size distribution and tortuosity, and corresponding process parameters such as capillary entry pressure and residual saturation.

Executive Summary

supersaturations. Here we report progress on all three topics.

software ADDICT, and (3) developed a new crystal growth model that is accurate over a wide range of

step propagation across crystal surfaces, (2) completed incorporating the COSMO solvent model into our

In the current funding period we have made advances on three fronts:

1. effect of impurity molecules on

Abstract

Executive Summary

This project sought to develop physically realistic models for atomization processes relevant to particle production, such as in spray-drying processes, with a focus on high viscosity and non-Newtonian fluid atomization. The goals of this work were to generate a spray database and to develop understanding and correlations for the accurate pilot-to-production scaleups. We divided the work to focus on two nozzle types: pressure-swirl, and two-fluid nozzles. The followings were achieved in the present reporting period.

Database was generated for two pilot scale nozzles.

• Droplet size distributions and near-nozzle images are obtained for the two pilot scale nozzles.
• Water (inviscid fluid), glycerin/water solution (Newtonian fluid), CMC/water solution (non-Newtonian polymeric fluid) are used as test fluids.
• Mass flow rate, spray angle and liquid sheet breakup lengths are measured for each case.
• Droplet size distributions are measured at 80mm downstream from the nozzle exit.
• Near nozzle images are taken from the nozzle exit to where the ligaments are formed.

Pressure-swirl nozzles: The following correlation is developed for high viscosity fluids.

For high viscosity and polymeric fluids, increasing the orifice diameter can result in a decrease in SMD.

The droplet size distribution of high viscosity and polymeric fluids has a bimodal distribution, with a minor and a major peaks (modes). A mixture of two lognormal distributions fits the volume distribution of all testing conditions. The minor peak (mode) of the size distribution represents the long tail of the volume distribution and it corresponds to droplets of less than 10 μm. These small droplets are mainly generated.

look at early formation stages of crystallization.

to study polymorph selectivity in different solvents or solvent mixtures as well as to

appropriate conditions, nucleation and growth on SAMs can function as a model system

interpretations. Overall, the two-part project clearly established, however, that under

were the basis for the designs of all experiments conducted and associated

Covid-19 shutdown). This may have led to variations in the solubility curves, which

that only a small amount of ultrapure ROY was available (in part as a result of the

crystallization experiments reported with ROY on SAMS were also plagued by the fact

likely caused by crystals falling off of the substrate into the solution. Furthermore, all

phenyl SAMs). The growth possibility dropped as supersaturation increased, which was

secondary polymorphs typically increased with increasing supersaturation (exception:

observed as a function of SAM chemistry. The possibility of occurrence of these

dominant polymorph, irrespective of SAM. However, secondary polymorphs were

with longer incubation time. When the solvent changed to toluene, Y was still the

(exception: phenyl SAMs). The growth possibility showed an overall increasing trend

polymorphs were not frequently observed, even at higher supersaturation levels

benzyl alcohol, Y was the dominant polymorph, irrespective of SAM. Secondary

applied for polymorph characterization of the resulting nucleated crystals. For ROY in

induced on the SAM surface by generating temperature jumps. Raman microscopy was

part of the project. SAMs were placed vertically in the solution and the nucleation was

conducted instead of the solvent evaporation-based crystallization employed in the first

control the level of supersaturation, crystallization experiments upon cooling were

in toluene and benzyl alcohol were determined. In this part of the project, in order to

on polymorph selection in concert with SAMs. To that end, solubility curves for ROY

chemistry, and both polar and nonpolar solvents were again chosen to study their effects

different terminal (omega) functional groups were used to control nucleating surface

(ROY), was chosen for in depth studies. Alkane-thiols based SAMs on gold with

organic model system, 5-methyl-2-[(2-nitrophenyl) amino]-3-thiophenecarbonitrile

other model compounds. In the second part of the project, to that end a more complex

formation stages of the crystallization of ACM and may be used to extend to studies of

suggested, that our methodologies are effective to gain insights into the earliest

possible existence of structural transformations at these early stages. These results

occurring. Further analysis and corroboration via additional data sets pointed to the

nucleation, we identified unusual shifts along scattering vector, q, of the earliest peaks

totally in-plane orientation. Studying crystallization of Form I by spontaneous

crystallographic orientation, directing the (002) planes from slightly out-of-plane to a

(trichloro(phenyl)silane) terminated SAM surface has a strong influence over

the substrate-solution interface than in the bulk above, and that a PTS

crystallization of Form II by seeded nucleation, we verified that crystals grow faster at

of Form I and II crystallization events of ACM under these conditions. Studying

Cornell’s High Energy Synchrotron Source (CHESS) to study the early formation stages

We then introduced time-resolved in situ wide-angle X-ray scattering (WAXS) at

surface chemistry and solvent conditions work together to control crystal polymorph.

degree of supersaturation. Under these conditions, we first demonstrated that both SAM

and growth were induced by simple solvent evaporation, i.e. without control of the

predominant crystal forms, Form I and Form II, was investigated. In this part, nucleation

part of the project, a pharmaceutical compound, acetaminophen (ACM), with two

chosen to study their effects on polymorph selection in concert with SAMs. In the first

used to control nucleating surface chemistry. Both polar and nonpolar solvents were

gold or silanes on glass, each with different terminal (omega) functional groups, were

in order to control crystal polymorphs. To that end, either alkane-thiol based SAMs on

monolayers (SAMs) in conjunction with varying solvents or solvent mixtures

mechanisms are highly desirable. In this project, we used a combination of selfassembled

methods that lead to an advanced understanding of early crystal formation pathways and

industries. Since a polymorph is determined at the early stages during crystallization,

of organic compounds is scientifically and technologically important to several

Understanding and control of crystallographic polymorphism and crystal habit

SUMMARY

Executive Summary

transport equations in order to supply quantitative prediction of segregation at process scale.

many of these segregation models and have set the stage for these models to be used in device-level

are applicable to density, size, shape, and cohesive segregation. We have experimentally validated

inherently-scalable, theoretical models based on rheologically-relevant dimensionless groups that

may be considered state-of-the-art, but, more importantly, we development several novel

than any models previously reported. Thus far we have demonstrated which models from the literature

a new way of structuring segregation rate models that make them inherently more scalable

novel view of the interplay between granular rheology and segregation, we aim to continue to develop

ultimately deduce the segregation rate (and validate the expressions). Moreover, by exploring a

model expressions that we are interested in testing with dramatically simplified experiments to

this balance between the rate of segregation and the perturbation rate, we can combine the

free surface granular flows in order to alter the steady-state distribution of particles. By achieving

is that we use flow perturbations to establish an “equilibrium” between segregation and mixing in

combined theoretical, computational, and experimental program. One unique aspect of our work

In this project, we seek to alleviate these two shortcomings of segregation research through a

• for validation purposes
• the significant dearth of validated scale-up studies for these models.

are (1) the inherent difficulty in measuring segregation rates (especially in an experimental setting)

practice. Two significant issues that hamper the applicability of models in industry, however,

Segregation model development holds promise for translation of academic research into industrial

will be measured to test the accuracy of their measurements.

measurements of both instruments at low stresses. The flow condition out of the hopper

not agree. A small hopper has been designed for gypsum powder based on the flow

the FT4 Powder Rheometer, however the measured values of unconfined yield stress do

cell and for the ball indentation method by sieve filling using an indentation attachment to

The reproducibility of flow measurement at low stresses is similar for the RST-XS.s shear

stresses as low as 100 Pa, however the variability increases for more free-flowing powders.

measurements with a coefficient of variation < 3% for the very cohesive titania at pre-shear

measurement in this range. The Schulze RST-XS.s provides unconfined yield stress

the intended applied stresses for titania DT51, and therefore does not provide a reliable

cell agree at moderate stresses, however below 2 kPa the FT4 shear cell does not achieve

Shear cell measurements using the FT4 shear cell and Schulze RST-XS.s low stress shear

than that of ball indentation and some shear cells.

the minimum consolidation stress that can yield a measurement with this method is greater

stresses the uniaxial compression test underestimates the unconfined yield stress, whilst

unconfined yield stress to be inferred from indentation measurements. However, at lower

moderate stresses, allowing constraint factor to be determined and ultimately for

Uniaxial compression measurements correlate with ball indentation measurements at

is followed.

packing state can be achieved by vertical consolidation when this bed preparation method

consistent across the radial direction of the powder bed. This indicates that a reproducible

excess powder heap away with an inclined blade, the flow resistance is found to be

consolidation stresses. Furthermore, by completely filling the powder bed and scraping the

indentation, whereas sieve-filling is able to produce a coefficient of variation < 3% at low

is found to provide a large coefficient of variation in the bed hardness measurement by

scraping away excess powder) prior to vertical consolidation. At low stresses, pre-shearing

and by blade conditioning, wire conditioning and sieve filling (with and without

approaches does not achieve this alone. Ball indentation measurements are assessed by preshearing

however the vertical consolidation applied in indentation and uniaxial compression

Shear cells pre-shear the sample in an effort to ensure a reproducible packing state,

shear cell methods.

measurement at low stresses is assessed here by ball indentation, uniaxial compression and

force measurement and (ii) the reproducibity of the loose packing state. Powder flow

measurement at low stresses is more challenging due to (i) the required resolution of the

the reliability and reproducibility of the measurement decreases. Powder flow

using shear cell measurements. As the consolidation stress applied to the powder is reduced,

with procedures developed for designing hoppers based on mass or funnel flow behaviour

The most widely established method for measuring powder flow is by shear cell analysis,

Measurement of powder flow behaviour is important for many powder handling operations.

xecutive Summary

way for a new production method for reproducible models of irregular powder beds.

deviation between experimental results and theoretical predictions observed, we paved the

artificial models to represent porous particle beds for liquid imbibition study. With some

considered. Thus, we 3D printed porous substrates building from simple to complex as

tortuosity in real-life powder beds or other industrial porous media usually are not fully

assumes a particle bed as an array of parallel capillaries. However, the complexity and

Wettability analysis for particulate materials has relied on the Washburn theory that

printed particles with complex geometry.

the first report of stress visualization and semi-quantification under low loads for 3D

material used in Polyjet printing was found to exhibit photoelastic properties. We presented

viable technique in understanding complex particle breakage behaviors. The Vero Clear

typically observed in Finite Element Simulations. Stress visualization also proves to be a

demonstrated the agglomerate strain distribution from the experiment in the same way

breakage from three dimensions using Digital Image Correlation (DIC). Preliminary results

In the fifth year, a more sophisticated approach was attempted to observe agglomerate

tuned by changing the liquid to powder saturation level.

Binder Jetting technique was used to produce agglomerates where the strength could be

over the agglomerate structure was plotted for the first time. In an additional study, the

tracking of individual particle position after agglomerate breakage. The strain distribution

properties. This was to provide feasible and accurate control on loading direction and better

process to produce 3D printed agglomerates with different color distributions and material

Building upon the initial findings, in the fourth year, colors were introduced in the printing

compressive load during the initial deformation of the agglomerate.

accurate predictions of the macroscopic breakage behavior and quantitatively predicted the

properties matching the 3D printed agglomerates. Qualitatively the DEM produced

simulated in EDEM using the Timoshenko Beam Bond Model (TBBM) with bond

tested under various standard breakage tests. Agglomerate deformation and breakage were

agglomerate design was systematically varied in terms of structure and bridge strength and

technique was used to print symmetrical or random agglomerate structures. The

the 3D printed structures and validating the results in the DEM simulations. The Polyjet

The first three years of the project investigated agglomerate breakage experimentally using

3. Agglomerate Flow and Segregation.

2. Agglomerate Disintegration and Dissolution.

• Agglomerate Breakage.

categories for a better understanding of agglomerate behavior:

printing techniques based on the properties required. The study was divided into three

Computer Aided Software’s (CAD) software’s, and printed using a wider range of 3D

overcome this barrier. Agglomerates were designed using various tools i.e. DEM and

new approach which involves 3D printing test agglomerates with “tunable” properties to

lack of suitable test particles that can be used to validate the models. This report presents a

One of the long term barriers to Discrete Element Modelling (DEM) of particulates is the

Executive Summary

Executive Summary

This report summarizes the main achievements of the three years of development of new crystallization technologies for improved crystal size and shape control during the crystallization process. The successful crystallization process and system design requires an interdisciplinary effort, which ranges from population balance model (PBM) development of the system concept, through efficient implementation of model equations to soft-sensor development, which is required for the model predictive control (MPC) design as well. This report gives a deeper insight into these interdisciplinary development efforts, which also highlights the achievable improvements enabled by the combination of process modeling, high performance process simulation and optimization.

In the first year we showed that the application of wet-milling during crystallization can significantly improve the process flexibility and attainable crystal size domain. The GPU acceleration halved the simulation time, which enabled faster optimization. In the same year we also developed a portable PBM solver for batch crystallization of 1D and high aspect ratio 2D particles with optional GPU acceleration, involving primary and secondary nucleation, crystal growth and dissolution. For model based control applications, a fast, approximate, geometrical model based CSD to CLD and aspect ratio distribution (ARD) transformation was also developed.

In the second year the concept of integrated crystallizer-wet mill process simulation and optimizing design has been extended to 2D high aspect ratio crystal case. Furthermore, using the portable PBM solvers developed in the first year, we designed and implemented a full PBM based nonlinear model predictive control (NMPC) system for the batch cooling crystallization of L-ascorbic acid. This NMPC, in addition to solved the barrier of computational time on an average PCs also used the FBRM CLD as a feedback device, for the first time in the literature.

In the third year, the optimal operation of integrated crystallizer-wet milling systems was generalized by discovering a repeating crystallization mechanism scheduling pattern in the optimal operating profiles regardless of the applied process model (1D or 2D), initial conditions (seeded/unseeded and seed loading) and the objective (smaller or larger desired crystal size). Furthermore, by improving the calculation performance of the portable GPU accelerated PBM solver, it was showed that the real time model based shape- control became real-time feasible. In addition, the iterative model based experimental design (IMED) was applied for a spherical agglomeration process, which work demonstrates how the experimental conditions can be design to minimize the number of experiments for precise parameter estimation.

Achieved deliverables

1. Development of a generic Matlab based software package for the simulation of 1D and 2D batch and continuous MSMSP crystallizers.
2. Quick design method for integrated batch crystallizer-wet mill systems.
3. Fast and approximate 2D CSD to CLD and ARD transformation development.
4. 1D NMPC development and implementation for L-ascorbic acid crystallization.
5. Real-time feasible 2D NMPC simulation.
6. Application of IMED for a spherical agglomeration process.

Executive Summary

The renewal IFPRI project revolves around utilizing 3D printing as a means to procure tuneable particles for the purpose of advancing particle technology. There are a number of advantages to this approach. Firstly, it provides accurate data for the particle properties which can be entered into a simulation, for comparison against the real experiment. This provides more precise results than the typical approach of measuring existing, non-identical particle properties. Secondly, an infinite number of experiments destructive in nature (i.e. breakage, dissolution) can be replicated to test a particle under the identical conditions.

In the previous IFPRI project, only agglomerate breakage was investigated and validated with a DEM simulation.

In year one of this project, the quasi-static compression tests of a random spherical agglomerate 3D printed in colour was carried out to investigate the distribution of strength. Soft, rigid and a hybrid mixture of both soft and rigid materials were used for the agglomerate bonds to determine the optimum bond material to prevent individual particle breakage, something which had not been previously observed. Further tests were undergone on agglomerates containing internal voids within the structure. X-ray tomography of our 3D printed particles as well as real granules were obtained for set up future projects.

In year two of this project, we report progress on 3D motion tracking of particle breakage, photoelastic polymer discs and 3D printed porous substrates for analysis of the wetting behaviour. All 3D printing files for the particle models have been uploaded onto our Thingiverse account. This report summarizes the progress made in the second year of the renewal project and provides a summary of the remaining work for the final year.