ARR - Annual Report
Proof of Concept
Reference: Sebastian Bindgen, Frank Bossler and Erin Koos “Defining structural transitions in capil- lary suspensions” Manuscript in Progress.
The structural properties of multiphase systems are essential to overall processability, functionality and acceptance among consumers. Therefore, it is crucial to understand the intrinsic connection between the microstructure of a material and the resulting rheological properties. Here, we demon- strate, how the transitions in the microstructural confirmations can be quantified and correlated to rheological measurements. We find methods from graph theory and thus from the mathematical study of networks, especially the clustering coefficient, to be a useful addition in accomplishing a link be- tween these two characterization methods. Our results, using capillary suspensions as a model system, show that the use of the clustering coefficient, in combination with the coordination number, is able to capture not only the agglomeration of particles, but also measures the formation of groups.
industry.
experimental results, and used to inform a `best practice' for the application of DEM in
number of benchmark systems are assessed and compared both to one another and to
(DEM) modelling to participate in a `round robin' in which their attempts to model a
This project brings together a cohort of companies who utilise discrete element method
1. Brief Summary of Project
Introduction to the Project
Executive Summary
For the year of 2019, an in-depth investigation was carried out into microstructural filter cake analysis. As planned comprehensive research involving of laboratory, modelling and simulation approaches were applied to characterize the filter cakes precisely. This was done for the case of hydrophobic and hydrophilic particles, forming hydrophobic and hydrophilic filter cakes.
As applying the commercial software approach for micro-CT analysis for filter cakes, we were limited due to the predefined functionality. To implement a plug and play analysis was not satisfying us to trust the output data fully. For this reason, we validate the filter cake characteristics based on four methods. One method is the laboratory approach in which filtrations were carried out in a standard VDI nutsch filter to measure pore size distribution, porosity and capillary pressure curves. In parallel CT scans of the same filter cakes were done to study cake characteristics based on the 3D reconstructed volumes by modelling and simulation. This is the imaging analysis approach, which itself is divided into three approaches.
The first image-based approach uses commercially available software only. In our case the software VGSTUDIO MAX 3.3 (subsequently called VGSTUDIO) was used. This method is very fast but the results deviate significantly from the laboratory data. The second image-based approach uses customized algorithms applied on the raw images. This approach is called customized code within this report. Using customized code showed good agreement with laboratory data. Nevertheless, this approach is time-consuming and requires advanced knowledge in image analysis and programming. The third method is to generate a pore network of the filter cake using laboratory data such as pore size distributions. Based on the generated pore network, simulations can be run to make prediction on e.g. desaturation of the filter cake (pore network modelling). For pore network modelling the open-source tool open PNM was used. The methodology and validation of each method is discussed in the report.
Executive Summary
In our second phase of the project (year two) we build upon the serendipitous finding made in phase one (year one): The significance of the new Pèclet number Pe = γ˙ d/√T extends beyond simply quantifying the relative importance of advection to diffusion, Akin to the definition of the Inertial number (I) and viscous number (Iv), Pe also represents the ratio of microscopic kinetic rearrangement timescale d/√T to the macroscopic shearing timescale (1/γ˙ ). Consequently, kinetic rearrangement—the fundamental driver of mixing—to the definition of the Inertial number (I) and viscous number (Iv), Pe also represents the ratio of microscopic kinetic rearrangement timescale d/√T to the macroscopic shearing is now controlled by the actual relative motion between particles, as opposed to assuming that the microscopic motion (so-called particle fall) is driven by pressure.
So building the new Pèclet-based rheology into a macroscopic-level process model of the rotating drum flow, could yield after suitable scaling, a dimensionless number that fully characterises the local mixing state of the granular assembly. And in terms of the IF- PRI project, the dimensionless number should facilitate robust scaling of the mixing state between different operating conditions.
However, before building the process model and dimensionless number upon a possible “house of cards” we embarked on an extensive testing process of the Pèclet-based rheology via simulation. Within the context of the DEM simulation framework, we treat the DEM outputs as “data” and apply coarse graining to yield the continuum-level stress and strain fields. Using > 200 simulation configurations spanning dry, wet, dense and dilute granular flows, we show that the Mohr-Coulomb friction coefficient µ = τ /σ, where τ is the shear stress and σ the pressure, varies linearly with √Pe for a variable yield stress ratio µs consistent with the different geometric configurations and flow regimes investigated.
µ = cµ√Pe, (1)
where the scaling parameter cµ is dependent on driving conditions, and hence the material contact friction coefficient µc.
The linear collapse of the data according to equation (1) spanned a wider range of the phase space than the µ (I)-rheology, viscoinertial rheology, non-local rheology and extended kinetic theory. We argue that cµ partly encodes the anisotropy of the tangential contact force network, and equation (1) relates this structure anisotropy to the stress ratio. Noting that these anisotropies are also signatures of the granular mixture state, we hypothesise that the most robust mixing rules will emerge from encoding the new Pèclet-based rheology into the formulation of the mixing rules.
Abstract
We report our recent results on the effect of impurities or ‘imposter molecules’ on the crystal growth process and developing mechanistic models describing the mechanisms by which impurities influence crystal growth. Impurities affect growth kinetics at the scale of kink attachment and detachment events, which are too fine to examine experimentally in real time. Thus, simulations are used to study the proposed mechanisms for growth inhibition/promotion and the results will be compared to experimental data available in the current literature. As a starting point, Kinetic Monte Carlo (KMC) simulations are utilized to simulate the time evolution of centrosymmetric organic crystal growth. Rare event rates are determined as functions of energetic barriers for desolvation and attachment/detachment works. Various mechanisms have been proposed to explain the growth-inhibiting effect of impurities, including step-pinning and spiral-pinning, which are described herein. These mechanisms will be incorporated into the KMC simulations, and compared to experimental values for validation. Once we have established effective working models, we will look to publish our results and ideally incorporate them into ADDICT3 (Advanced Design and Development of Industrial Crystallization Technology - version 3), an engineering tool which predicts relative growth rates and crystal morphology of solution-grown faceted crystals [1].
Additionally, we report on some new developments in ADDICT3 to improve the workflow during conceptual design of crystalline solid processes. ADDICT3 is capable of predicting the shape and morphology of crystalline particles using only the following input data, (1) crystal structure, (2) atom-atom force field, and (3) specified solvent and growth conditions. The key outputs are the steady-state growth shape and morphology of the crystals, and the “shape triangle” which gives graphical information about the influence of design variables (temperature, supersaturation, solvent) on the resulting shape and morphology of the crystal. More detailed outputs are also provided (e.g., periodic bond chain networks, growth spiral shapes, etc.) for use by the more sophisticated user. The tool can be used alone or in combination with other tools that have been developed recently to aid in the design of crystallization processes.
Executive Summary
Understanding and control of crystallographic polymorphism and crystal habit of organic compounds is scientifically and technologically important to several industries. Since a polymorph is determined at the early stages in crystallization, methods that lead to an advanced understanding of early crystal formation pathways and mechanisms are highly desirable. In this work, we have introduced time-resolved in situ wide-angle X-ray scattering (WAXS) at Cornell’s High Energy Synchrotron Source (CHESS) to study the early formation stages of the Form I and II crystallization events of a pharmaceutical compound, acetaminophen (ACM). Studies were informed by results from earlier investigations that both the surface chemistry of self-assembled monolayers (SAMs) as well as solvent conditions work together to control crystal polymorph. Studying crystallization of Form II by seeded nucleation, we verified that crystals grow faster at the substrate-solution interface than in the bulk above, and that the PTS (trichloro(phenyl) silane) SAM surface is taking over in controlling crystal growth by directing the (002) planes from slightly out-of-plane to a totally in-plane orientation. Studying crystallization of Form I by spontaneous nucleation, we identified unusual shifts along the scattering vector, q, of the earliest peak occurring in the in-situ scattering patterns. Further analysis and corroboration by other data sets pointed to the possible existence of structural transformations at these early crystal formation stages. Our results indicate that our methodologies are effective to gain insights into the earliest formation stages of the crystallization of ACM and are now being extended to other model compounds.
The reproducibility of flow measurement at low stresses is similar for the RST-XS.s shear cell and for the ball indentation method by sieve filling using an indentation attachment to the FT4 Powder Rheometer, however the measured values of unconfined yield stress do not agree. A small hopper will be designed based on the flow measurements of both instruments at low stresses. The critical outlet diameter for flow initiation will be experimentally determined and compared to the dimensions predicted by the flow measurements of these two instruments. To minimise the powder quantity, gypsum powder is proposed for this investigation due to its high bulk density and cohesive flow behaviour.
Shear cell measurements using the FT4 shear cell and Schulze RST-XS.s low stress shear cell agree at moderate stresses, however below 2 kPa the FT4 shear cell does not achieve the intended applied stresses for titania DT51, and therefore does not provide a reliable measurement in this range. The Schulze RST-XS.s provides unconfined yield stress measurements < 5% for the very cohesive titania at pre-shear stresses as low as 100 Pa, however the variability increaseses for more free-flowing powders.
Powder flow measurement at low stresses is more challenging due to (i) the required resolution of the force measurement and (ii) the reproducibity of the loose packing state. Shear cells pre-shear the sample in an effort to ensure a reproducible packing state, however the vertical consolidation applied in indentation and uniaxial compression approaches does not achieve this alone. Ball indentation measurements are assessed by pre-shearing and by blade conditioning, wire conditioning and sieve filling prior to vertical consolidation. At low stresses, pre-shearing is found to provide a large coefficient of variation in the bed hardness measurement by indentation, whereas sieve-filling is able to produce a coefficient of variation < 5% at low consolidation stresses. Uniaxial compression measurements correlate with ball indentation measurements at moderate stresses, allowing constraint factor to be determined and ultimately for unconfined yield stress to be inferred from indentation measurements. However, at lower stresses the uniaxial compression test underestimates the unconfined yield stress.
Measurement of powder flow behaviour is important for many powder handling operations. The most widely established method for measuring powder flow is by shear cell analysis, with procedures developed for designing hoppers based on mass or funnel flow behaviour using shear cell measurements. As the consolidation stress applied to the powder is reduced, the reliability and reproducibility of the measurement decreases. Powder flow measurement at low stresses is assessed here by ball indentation, uniaxial compression and shear cell methods.
Executive Summary
Figure 1. Take home message from the first year.
A first statistical classification was established for thirty powders. Four categories (green, yellow, orange and red) corresponding to various reconstituabilities were identified and summarized in the take home message (Figure 1). Long wetting time was associated to high particle surface hydrophobicity, small particle size, high protein and fat contents in the powder bulk. Long reconstitution time was linked to the powder manufacturing process (grinding) and low sugar content in the powder bulk.
- Green group: short reconstitution time and short wetting time,
- Yellow group: short reconstitution time and long wetting time,
- Orange group: long reconstitution time and short wetting time,
- Red group: long reconstitution time and long wetting time.
(3) First statistical correlations between the various powder characteristics and reconstituability ranking
(2) For selected powders: reconstitution kinetics in different conditions of temperature, stirring, etc.
(1) Powder classification according to their reconstitution behavior
Achieved deliverables:
The first year of the PhD work deals with the systemic physicochemical analysis of powders and their classification according to their reconstitution ability. It will be achieved at the end of January 2020. The PhD work has met the three stated deliverables for thirty powders.
(PhD duration: 1st February 2019 – 31th January 2022)
Executive Summary until the tenth month
EXECUTIVE SUMMARY
In the field of granular rheology, one of the most promising advances of the past decade has been the development of various nonlocal rheologies [1–7]. These constitutive models hold the promise of permitting the determination of a small number of empirical parameters for a particular set of particles, which then can be used to predict flow fields and stresses over a large range of intermittent, creeping, quasi-static, and intermediate flows. In order for these models to be useful, the aim is to make a set of flow measurements for a set of particles in one geometry, and then determine the constitutive parameters for use in predicting flows in other geometries (for the same particles). Doing this requires a quantitative understanding of which properties are set by both the particle properties, and the boundary conditions at the walls.
In Years 1-3, we established that NLR successfully models granular flows across different packing densities, particle sizes and shapes, and shear rates, using just 3 constitutive properties (A, b, µs), but that we must know the amount of slip at the wall from geometry-dependent measurements. In Years 4-6 of this project, we aim to address current shortcomings in how to calibrate and apply NLR to real granular systems. We aim to establish that, for a given set of particles, can we (1) make flow measurements in one geometry which determine the constitutive parameters and (2) use these parameters to predict flows in other geometries. Thus, our current work focuses on separating which flow properties are set by the particle properties, versus those set by the wall properties.
During this first year of the renewal grant, we have observed that boundary roughness strongly controls both the flow profile v(r) and shear rate profile γ˙ (r), particularly as measured at the outer wall. This is also the region of the flow most sensitive to nonlocal effects. We also observed that the pressure P , and therefore the stress ratio µ(r), are affected by the roughness of the wall. All of this is expected, and we can now begin a quantitative understanding of these effects during Year 5. We have done significant development work on photoelastic methods to prepare for these studies. In addition, we continued working on the IFPRI-funded collaboration with Nathalie Vriend.
This resulted in measurements taken at faster flow rates than are accessible in this apparatus (I 1), and resulted in a published paper [21] that included NCSU authors. Along with our other work during Years 1-3, this paper emphasizes the importance of understanding the timescales over which the force chains fluctuate. In performing the next year’s work, will utilize photoelastic methods (both statics and dynamics) as well as our older particle-tracking methods to meet the three aims.
Executive Summary
This project seeks to develop physically realistic models for spray-drying as used in industry with a focus on high viscosity and non-Newtonian fluid atomization. The goal of this work is to develop understanding and correlations for the accurate pilot-to-production scaleup of spray-drying. We divide the work to focus on two nozzle types: pressure-swirl, and two-fluid.
A testing facility is setup for experiments with swirl nozzle. Various types of pressure swirl nozzle with different scales are selected to be tested with different fluids at relatively high pressure. An imaging system is used to take near-nozzle images to analyze the atomization process. It is also used to measure the droplet size downstream and report number distribution and SMD at various radial positions. An atomization process is proposed by observing near-nozzle images. Physics model is being developed and its predictions will be compared with all the experiment results for further improvements.
Preliminary tests of two-fluid nozzles reveal the similarity to single-droplet breakup, thus detailed experimentation and analysis on single droplet atomization by an air jet was conducted. A working hypothesis that the rate of droplet deformation determines the mode of droplet breakup is being studied, and models for the underlying breakup processes have been compared to the experiments with reasonable agreement. The next phase for this branch of the project will be to apply the developed models for single-droplet breakup to the two-fluid atomization geometry.