ARR - Annual Report

Executive Summary

2019.

Compacting GmbH. It is expected that the system will be ready for some preliminary test in Q2 of

Furthermore, a forced filling system is currently under construction in collaboration with Fette

obtained but further data analysis is required. These results will be reported at IFPRI 2019 AGM.

constructed and a systematical experimental work was perfomed. Some interesting results were

In addition, significant progress was also made in exploring suction filling, an experimental rig was

air sensitivity index increase, while it decreases with the increase of specific energy and cohesion.

filling velocity increases proportionally as the mean particle size, flow function, air permeability and

as cohesion, flowability, average particle size and air sensitivity index. In particular, the critical

filling velocity. It was found that the critical filling velocity is strongly dependent on such properties

properties were examined. The efficiency of die filling is evaluated using the concept of critical

material characteristics (e.g. particle size distribution, sphericity and morphology) and flow

this system, die filling behaviours of 7 commonly used pharmaceutical excipients with various

filling occurs when the die passes through a stationary shoe positioned above the die table. Using

rectangular die. The die table can rotate at an equivalent translational velocity of up to 1.5 m/s. The

tablet press. The system consists of a round die table of 500 mm diameter, equipped with a

A model rotary die filling system was developed to mimic the die filling process in a typical rotary

Ms. Zakhvatayeva. It primarily covers a comprehensive study of rotary die filling.

This report summarizes the work performed during the last 12 month primarily by the project student,

Executive Summary

constraint factor increases with an increase in rolling friction coefficient or interface energy.

suggest that the constraint factor remains constant across all stress levels, and that

the addition of coarse material and decrease with the addition of fines. DEM simulations

The constraint factor of bi-disperse, cohesive glass beads has been shown to decrease with

long as they cover a sufficiently wide range.

applied in a shear test are shown to have limited influence on the generated yield locus, so

insufficient to overcome the influence of prior handling. The precise normal stresses

driven by variability in the pre-shear shear stress, thus implying the conditioning is

at lower stresses the variability in measured yield data increases; this appears to be largely

normal stress data being unavailable. Repeated testing in the Schulze shear cell shows that

titania powders. A direct comparison could not be made with the Schulze shear cell due to

been shown to apply normal stresses noticeably greater than the target stresses for these

shear cell and the Schulze RST.XS.s shear cell. At lower pre-shear stresses the FT4 has

The flow behaviour of various titania grades is found to be largely similar between the FT4

universal for all powders.

measurements to be made at lower stresses, though the optimal lid design is unlikely to be

in the shear plane. Reducing the vane height of the shear cell lid allows indentation

to critically consolidated beds since an insufficient fraction of the exposed bed is located

consolidated beds. It is shown that at lower stresses it is challenging to apply indentation

condition, with critically consolidated beds providing a greater hardness than vertically

Indentation hardness measurements have been shown to be dependent on the loading

investigated using the Discrete Element Method (DEM).

experimentally, and the influence of a broader range of particle properties has been

size distribution and interface energy on constraint factor have been assessed

for several grades of titania with two different shear cells. Furthermore, the influence of

has been assessed in ball indentation, and shear cell measurements have been carried out

used to measure flowability of titania under a range of stresses. The bed preparation method

inconsistent. The ball indentation, uniaxial compression and shear cell methods have been

Measurement of powder flowability under low stress conditions is often unreliable or

Executive Summary This report summarizes the main achievements of the second year’s effort 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 report 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 this year, we applied the same principles, but we investigate the effect of milling in addition to crystal size also on the crystal shape. Since the computational cost associated to numerically solve the model equations is a power-law function of number of dimensions, the application of 2D PBMs for this purpose is not feasible without GPU acceleration, which in this case brought ~1.5 order of magnitude speedup. The first, full PBM based nonlinear model predictive control (NMPC) was implemented for the batch crystallization of L-ascorbic acid. The NMPC showed good control behavior, produced significantly better crystals than the direct nucleation control under considerably shorter batch time. The novel aspect of this NMPC was that it applied the fast, approximate CSDCLD transformation, that was presented in the previous report. The development work of the analogue control system for simultaneous size and shape control has been started. A major challenge, associated to the calculation time was solved.

Realized deliverables

1. Model development, simulation and preliminary optimization of an integrated batch crystallizer-wet mill system for bivariate crystal size distribution control
2. Implementation of a full PBM based NMPC for crystal size distribution control
3. Preliminary development work of a real-time feasible full 2D PBM based adaptive predictive simultaneous crystal size and shape control algorithm

structures collected from a lab scaled spray dryer.

Structures obtained from the single droplet drying experiments showed high similarities to

The differences in these drying metrics have been related to key material properties.

final particle size, moisture content at locking were collated at different drying conditions.

encountered above boiling. Several drying metrics including relative size at locking, relative

further understand the differences in drying behaviour and morphology evolution route

conditions collected from the single droplet drying rig have been extensively investigated to

deflation, inflation and puffing. Drying kinetics and morphology time-series at different drying

different drying mechanisms, boiling without any inflation/deflation cycles, inflation and

concentrations using a filament single droplet drying rig. The three materials showed three

and drying kinetics have been investigated across a range of air temperatures and initial solid

sucrose, sodium silicate and Hydroxypropyl Methylcellulose (HPMC). The drying behaviour

and morphology evolution. Three skin forming materials were chosen for this study namely,

The aim of this study is to investigate the effects driven by boiling on the drying behaviour

Wael Ebrahim

above boiling point

An experimental investigation of the drying mechanisms of single droplets

The solution of this model is currently underway.

expression for the expansion of a single centralized bubble within a liquid droplet was derived.

transfer on the expansion of a single bubble in an infinite liquid media was conducted and an

to the asymmetric pressure and velocity field. An initial assessment of the influence of mass

An off-centre bubble was also investigated and a self-centering behaviour was observed due

oscillations which decayed with time. The rate of decay increasing with increasing viscosity.

dynamics of a initially over pressured bubble, behaved as anticipated showing bubble

constructed and solved to allow for asymmetry to be investigated. The 2-D model, of the

spherically symmetric 1-D equations were derived and solved and a 2-D model was also

derived from the mass and the momentum balances and the Navier-Stokes equation. Both

The governing equations for the oscillations dynamics of the bubble and the droplet are

within a droplet, which is termed as ‘bubble-droplet system’ in this report, are investigated.

The effects of mass transfer on the oscillation dynamics of a single bubble centrally located

Tien Nguyen

BUBBLE DYNAMICS INSIDE A DROPLET

Executive Summary

hydrates/solvates, and cocrystals crystallized from solution.

we plan to continue with the remaining modifications required to implement models for organic salts,

ADDICT v3.0 against molecular crystals with multiple molecules in the asymmetric unit. Following this,

Upon completion of this new architecture, we will demonstrate improved functionality by testing

co habit predictions of organic salts, cocrystals and solvates grown from solution.

quires a complete rewrite of the existing codes, these modifications lay the groundwork for rapid in sili-

lographic complexity. We are currently implementing this redesign within the software. Although it re-

of the solid-state interactions that is independent of the asymmetric unit and applicable to any crystal-

solvates and cocrystals). With our redesigned input preparation architecture, we can build a description

sponds to the growth unit itself, many systems do not satisfy this criterion (including all organic salts,

from the asymmetric unit. While this approach is feasible for systems where the asymmetric unit corre-

architecture for acquiring this information is to apply crystallographic operations to generate a unit cell

growth models is to calculate and organize solid-state interactions between growth units. The typical

salts, cocrystals and solvates. A necessary, but not sufficient, condition to enact mechanistic crystal

shape-prediction design aid that is applicable to all crystalline solids, from organic molecules to organic

vanced Design and Development of Industrial Crystallization Technology). Our goal is to produce a

important solid forms, and especially drug substances. We call our software design aid Addict (Ad-

morphology software tool in order to generalize the methodology to a much broader class of industrially

In the first year of this project we have used IFPRI funds to redesign and rewrite our crystal growth and

Executive Summary

relations based directly upon PEPT-derived stress computations.

in greater detail in the next phase of the project wherein we formulate mixing scaling

timescale to the macroscopic shearing timescale more naturally. This idea will be explored

like the Inertial number, also appears to describe the ratio of the microscopic rearrangement

extend beyond simply quantifying the relative importance of advection and di.usion, and

number that is computed directly from the PEPT data. Interestingly, it’s validity seems to

an alternate definition. In this regard we propose a granular temperature-dependent Péclet

The limitations in the standard definition of the Péclet number precipitated the need for

mills) that operate in the high Froude regimes.

the data. These findings have significant implications for industrial systems (like tumbling

cascading and cataracting) the Reverse Brazil Nut E.ect (RBNE) is clearly evident from

the Brazil Nut E.ect (BNE) at low Froude numbers; however, at higher Froude numbers

The initial analyses of PEPT data clearly show radial segregation by size consistent with

developed flow conditions.

assumption, we extract the bed shape, solids fraction and kinematics for steady, fully

of representative radio-labelled beads (the tracer) into Eulerian fields under the ergodic

and cataracting Froude regime. After converting the measured Lagrangian trajectories

matrix spanned four fill fractions and seven drum rotation rates across the cascading

plastic beads) within a laboratory rotating drum fitted with lifter bars. The experimental

was used to measure the 3D trajectory of a binary mixture (3mm and 5mm diameter

In our first phase of the project (year one) Positron Emission Particle Tracking (PEPT)

many investigators that ultimately lead to a restricted mechanistic interpretation.

is not well understood. A further limitation relates to the low Froude regimes explored by

known to exhibit axial and radial mixing; however, their interplay for optimal performance

system encountered in the mineral processing, food and pharmaceutical industry is well

the high shear zones where most of the energy is dissipated. The canonical rotating drum

processes is the inability to control the relative mix of grinding balls and small rocks in

ingredient and excipient. Central to the low efficiency (< 5%) reported in mineral grinding

industry, the e.ectiveness of the drug is directly related to the mixture of active

Mixing (and segregation) play a vital role in several industrial processes. In the pharmaceutical

Executive Summary

In the first part of this report, we describe a series of simulations studying the effect of hydrodynamics on gelation dynamics, gel structure and gel sedimenta- tion. We find that hydrodynamics affects strongly the dynamics of gelation, but has little effect on gel structure. In the second, experimental, part, we first bring to a conclusion the previous four years of work on a model system consisting of sterically-stabilised polymethylmethacrylate particles forming a gel due to the depletion attraction induced by non-adsorbing polystyrene polymers. One of the main results of our previous work was that the curved meniscus had a major effect in inducing the gravitational collapse of these gels. Here, we directly manipulate the curvature of the gel meniscus, and show that curvature of either sign speeds up collapse. We conclude by reporting preliminary experiments on a new sys- tem consisting of large, repulsive particles dispersed in a background gel of small attractive particles. The rheology of this system, which models a range of indus- trial formulations, is surprisingly complex. We find that as we change the relative amount of large spheres, the system can display time or history dependence.

Executive Summary

Milling is commonly deployed in many industrial sectors for intended particle size reduction. In this project, we aim to develop a robust methodology to link material grindability with particle dynamics in a mill in order to provide an innovative step-change in mill fingerprinting and optimization. This involves characterizing the stressing events that prevail in a milling operation and establishing material grindability in the context of the stressing events. The material grindability will require a detailed study of the fundamental fracture and breakage mechanisms of individual particles under different loading regimes, and how they relate to the mechanical properties and the final size distribution. This will provide the fundamental scientific basis for developing appropriate grindability measure capable of analysing particle breakage subjected to particle impact, compression, and shear etc. pertaining to a milling process, which in turn will provide the basis for an improved particle breakage model calibrated against the defined grindability.

The centrifugal impact pin mill has been selected to be studied for this project, in collaboration with Hosokawa Micron Ltd. UPZ100 pin mill experiments with varying rotary speeds and feed rates were reported and analyzed in the past reports. The work performed in Year 5 of the project is to develop a coupling framework between discrete element method (DEM) and population balance model (PBM) to predict the product size distribution of milling experiments. At the particle scale, DEM simulations were performed to understand the fundamentals of the particle dynamic and stressing conditions inside the mills. Variables in the PBM kernel were classified into material dependent parameters and mill operation dependent parameters. The impact velocity distributions obtained through DEM simulations were utilized to inform the mill operation dependent parameters of PBM at the process scale. The remaining parameters of PBM, i.e. material dependent parameters, were estimated based on the milling test at 12000 RPM. The resulting DEM-PBM coupled model is then used to predict the milling results for the other three rotary speeds to validate the proposed DEM-PBM model. A good agreement between the tests and the predictions of product size distribution has been achieved, which indicates the potential application of the proposed DEM-PBM multiscale method for scale-up and optimization of milling processes. The follow-on work will focus on further improving the material dependent parameters evaluation and studying the particle breakage mechanism using the Edinburgh bond DEM model.

Executive Summary

During this reporting year, we focused on the following activities:

• Parametric studies were conducted that elucidate the effect of the differences in material properties of two components in a mixture on the strength of the mixture and some reasons that there is a deviation from the rule of mixtures.

• Appropriate definitions of damage were derived in order to understand the evolution of the strength of a compacted mixture during the compaction/unloading/ejection sequence.
• The issue of damage directionality become obvious in a parallel leveraged project. We present here only a single yet important relevant simulation. Detailed presentation of this topic will be done in the final report.

Executive Summary

The understanding and control of crystallographic polymorphism and crystal habit of organic as well as inorganic compounds is scientifically and technologically important to a number of industries. To date, however, the experimental control of polymorphs (crystalline solids with different arrangements of the same constituents) is difficult. Since a polymorph is determined at the nucleation of a crystal, methods that lead to an advanced understanding of early crystal formation pathways and mechanisms are highly desirable. Towards this aim, in this project we employ arrays of self-assembled monolayers (SAMs).

Self-assembled monolayers (SAMs) are well-defined surfaces that can be used to study the relationship between the nucleation event and the final polymorph selection. Furthermore, by tuning the substrate-crystal interface energy, potentially crystalline order of SAMs can promote the nucleation of polymorphs not accessible via solution methods. It is these two advantages, i.e. the establishment of scientific correlations between nucleation and observed polymorph and access to polymorphs not accessible via solution methods, that have led us in this project to choose heterogeneous surface nucleation via SAMs as the primary means to study polymorph selection.

In the second-year of work, we examined 1-undecanethiol (UDT) and 11-mercapto-1-undecanol (MUOH) SAM chemistries on gold, and trichloro(octadecyl)silane (OTS) and trichloro(phenyl)silane (PTS) on oxide bearing silicon substrates in the presence of various solvent systems to investigate their ability to influence the nucleation, polymorph selection and crystal growth of acetaminophen (ACM). We found that for evaporating solvent from a single droplet on SAMs, both solvent(s) and SAM substrate work together to control crystal polymorph selection. On hydrophobic surfaces (UDT, OTS, PTS), use of pure solvents resulted in ACM form I (monoclinic), while a mixture of water and dioxane produced form II (orthorhombic). In addition to polymorph selection, under these conditions we found that for form II different SAM surface chemistries influence crystal orientation. Finally, by introducing a doctor-blading process, for first experiments of PTS SAMs on a silicon waver, polymorph selectivity could be achieved varying the solvent from 1,4-dioxane to ethanol. This opens the door to similar experiments at the Cornell High Energy Synchrotron Source (CHESS) in the third-year period. By studying at CHESS the known relationships between the structure of the crystal on one side, and the nucleating surface and conditions (quiescent versus shear) on the other, we hope to gain insights into the early formation pathways of crystallographic polymorphs.