Particle Formation

of granule structures.

measurement will be combined on an significant set of granules, representing a variation

In the last part of this first period the two aspects of structure description and property

between structural and physical measures is the next step that was performed.

The combination of both kinds of measures to evaluate if there are any significant correlations

between differently structured granules.

The structural as well as the physical parameters provide results that are suitable to distinguish

granules are dispersed in an optical glass cuvette and measured continuously.

measurement of particles size distribution over time in a laser diffraction system. The

beaker test. The second method to measure dissolution behaviour contains the

of granules containing salt within the binder phase can be determined in a standard

Regarding dissolution and disintegration behaviour of the structured granules, the conductivity

positions of granules on a sample container and crushes these granules consecutively.

single granules has to be measured. An experimental set-up was designed that scans

To achieve statistically relevant results for the single particle crushing, a high number of

behaviour by conductivity measurements and online particle sizing respectively.

mechanical strength is measured by single particle crushing and the dissolution or disintegration

For the determination of physical properties three different methods were chosen. The

and covariance function.

were evaluated include volume and surface fraction, star volume, chord length distribution

and used for further calculations of structure descriptors. The structure descriptors that

Three dimensional X-ray micro tomography images of the model granules were recorded

coarse limestone particles was chosen and combined with a polyethylene glycol binder.

to achieve different internal structures. A bimodal particles size distribution of fine and

Model granules were generated containing different size distributions of primary particles

correlation to structure parameters.

this year was about the determination of physical parameters of model granules and their

While the last period of work was focused on the implementation of structure descriptors

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.

Executive Summary

In the immediate future, work within this project will focus on disintegration model sensitivity analysis, model calibration and validation. The disintegration model will then be linked to an existing model for high shear granulation, and the process of inverse problem solving for the linked process-product model initiated.

Initial experimental method development is also presented. In this work, the relationship between granulation process parameters, internal granule structure and granule dissolution kinetics was studied. The results demonstrate the important role of granule porosity in granule dissolution, and provide important steer for the development of validation experiments for granule disintegration.

A preliminary parametric sensitivity analysis was conducted for the unified granule disintegration model, which investigated the effects of initial granule size, disintegrant particle size, initial granule porosity and disintegrant absorption ratios on granule size distributions and size evolution. The model was found to be sensitive to all four parameters, with faster disintegration promoted by increased granule size, decreased disintegrant particle size, lower initial porosity and increased absorption ratios.

This annual report summarises key progress toward the project aims made over the past 12 months. In particular, a mechanistic single granule model has been developed, which includes rates processes such as liquid penetration, liquid absorbance, swelling and stress build-up. This single granule model has been coupled with a population balance model, to enable the modelling of swelling driven breakage for populations of granules.

The aim of this project is to demonstrate the ability to link process and product models for wet granulation processes, and to initiate solution of the inverse problem, i.e. to determine the required process parameters and material properties to provide desired granule performance. A case study of granule disintegration/dispersion has been chosen as the product performance model for this work.

1. the lack of mechanistic models which relate granule structure and material properties to granule properties, and
2. the cooperative development of linked process and product models, to ensure essential parameters are modelled throughout the process and product.

While these performance criteria are typically the driving reason to granulate in the first place, we are still unable to confidently design granulation processes which optimise for desired product performance. There are several reasons for this, but two of the largest hurdles to overcome are:

The vast majority of granular products are granulated to produce some desired function and performance. This function may be improved flow, reduced dustiness, specific granule strengths and attrition resistance, or could be specific criteria for dispersion or dissolution.

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

Abstract

The description of ‘smart’ is applied to a variety of systems that are capable of reacting to a change in environment and providing a unique response. The scope of this review is focused on environmentally responsive smart particles, with a secondary emphasis on in-situ sensor particles for in-process characterization. This review provides information on particles developed to ‘smartly’ respond to mechanical, thermal, chemical, biological, electromagnetic, magnetic, or electrical stimuli. The review covers multiple aspects of “smart” particles, including their applications, mode of action, preparation methods, and fate. Smart particles used for commercial and experimental applications are included.

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 project seeks 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 are to generate a spray database and to develop understanding and correlations for the accurate pilot-to-production scaleups. We have divided the work to focus on two nozzle types: pressure-swirl, and two-fluid nozzles. The followings are achieved in the present reporting period.

Pressure‐swirl:

• Completed testing of Fine Spray nozzles with several different fluids with varying viscosities. The droplet size distributions, closeup images of the atomization zone, and other statistics are obtained and reported.
• A correlation for SMD has been found using pressure-based Weber number and Reynolds number.
• The above correlation does not work well for sprays formed by small orifice nozzles using high viscosity fluid (60 cps). This was found to be due to change in the atomization mechanism for such cases. Atomizing very high viscosity fluid (60 cps) with small orifice nozzles may result in very large SMD, indicating a poor atomization quality. This is due to significant pressure loss across small orifices using high viscosity liquids.
• The above correlation has also been tested with experimental data by other researchers that have used different nozzle designs.

Executive Summary

driven model for granulation.

for granulation, and the application of inverse methods to create a product performance

incorporates the development, validation and integration of process and product models

A research plan for the remaining three years of the project is presented. This plan

characterisation of feed materials and granular products.

models, however this initial experimentation has emphasised the need for careful

will take place after critical decisions are made on choices of product and process

development for granulation and granule dissolution. Full methodology development

Also presented within this report is preliminary experimental methodology

absence of disintegration.

dispersion model, and the second a model for drug dissolution from granules in the

two potential product models presented here for use in this project. The first is a granule

available to describe these mechanisms. This review has informed the development of

disintegration and dissolution, and also includes a review of the mathematical models

review incorporates the current state of knowledge on the mechanisms of compact

behaviour, culminating in the literature review presented in this report. This literature

survey of the literature on granule, tablet and compact disintegration and dissolution

of Granular Products. Focus has been placed in this first year of the project on a critical

This report provides a summary of the progress of the project Model Assisted Design

address this need.

product models to enable performance driven process design. The aim of this project to

develop improved performance models for granular products, and to link these with

models have received less attention than process models, and there is a clear need to

not describe the performance of the products being produced. Product performance

outputs of these models are typically limited to one or two particle attributes, and do

processes. This is an exciting and welcome development for the field, however the

process design. Process models are increasingly being developed and used for these

and laborious experimentation, due to a lack of knowledge and predictive tools for

Traditionally, the design and scale-up of granulation processes has involved expensive

tool that predicts relative growth rates and crystal morphology of solution-grown faceted crystals [5].

into ADDICT (Advanced Design and Development of Industrial Crystallization Technology), an engineering

in the literature for validation. Once we have established effective models, we will look to incorporate them

considered in the context of KMC simulations for model development, and compared to experimental values

the growth-inhibiting effect of impurities, such as step pinning and spiral pinning. These mechanisms are

for desolvation and attachment/detachment works. Various mechanisms have been proposed to explain

centrosymmetric organic crystal growth. Rare event rates are determined as functions of energetic barriers

for growth inhibition. We employ Kinetic Monte Carlo (KMC) methods to simulate the time evolution of

ne to examine experimentally in real time. Thus, we use simulations to study the proposed mechanisms

Impurities affect growth kinetics at the scale of kink attachment and detachment events, which are too

and hence affect crystal morphology and size.

process and to develop theoretical models for the mechanisms by which impurities influence crystal growth

The goal of this project is to investigate the effect of impurities or `imposter molecules' on the crystal growth

are more representative of realistic conditions. One such non-ideality involves the presence of impurities.

equal surfaces). There is interest in studying crystal systems in which non-idealities are introduced, as these

systems i.e., Kossel crystals with a single centrosymmetric growth unit (simple cubic single molecules with

mechanistic modelling of crystallization. Crystallization of organic molecules is well understood for ideal

Given the ubiquity of crystal growth in industrial processes, there is substantial demand for predictive and

such as OLEDs [3] and for altering the impact sensitivity of energetic materials such as RDX and HMX [4].

crystallization is desirable for varying electrical and optical properties in the field of electronic materials

importance for developing catalysts with tailored surfaces to maximize active sites [2]. Furthermore, tuning

to selectively formulate specific crystal habits for optimal bioperformance [1]. Crystal engineering is also of

a structured solid state. Pharmaceutical companies often crystallize APIs in the form of organic molecules

Crystallization is commonly used in industrial processes to convert solute molecules dissolved in solvent to

Abstract