FRR - Final Report

This concluding report for the grant described the development of software and instrumentation to obtain information for understanding product formulation, notably for pastes, aerated pastes, foams and flowing suspensions.

Of particular note and novelty are the development of:

• An electrically-based system for in-situ monitoring of paste extrusion from which information on homogeneity of the product can be inferred in the barrel and die;
• software for interpreting conductivity data for high gas content foams to deduce the cellular density and, from temporal information, likely cell size;
• methods for using x-ray tomography to measure paste homogeneity (on ex-situ samples) to complement the above methods;
• methods for characterising sediments through 3d structure mapping that can then be used as inputs to lattice Boltzmann simulations to predict fluid permeability directly;
• very high speed electrical impedance tomographic (EIT) imaging methods that can be used to visualise flowing particulate suspensions up to 20 m/s and from which voxel-voxel correlation methods can yield angular, axial and tangential velocities directly;
• use of the EIT methods for quantitative measurement of fluid dynamics for comparison with or integration into process based models for fluid flow or computational fluids dynamic predictions.

Examples of these developments and background work are described and cited in this report together with support appendices.

R A Williams

Executive summary

The contractor (Fumio Saito) has conducted his group to investigate his work on mechanochemistry of materials supported financially by IFPRI Inc. for six-years since 1998 to 2003. The first year’s work has been devoted to study the effect of dry grinding of a CaO-SiO2 mixture on synthesis of para-wollastonite by heating, followed by synthesis of tricalcium aluminate hydrate (3CaO·Al2O3·6H2O (C3AH6)) by dry grinding a mixture of calcium hydroxide and boehmite. He has also dealt with mechanochemical (MC) direct synthesis of CaTiO3 from a CaO-TiO2 mixture and its high-resolution transmission electron microscope (HR-TEM) observation. These were the work done in 1998, and, the main results are as below:

The contractor has attempted to synthesize CaTiO3 from a mixture of TiO2 and CaO. It has been known that there are three polymorphs in TiO2, i.e., anatase (TiO2, tetragonal), rutile (TiO2, tetragonal) and brookite (TiO2, orthorhombic). The use of anatase leads to MC synthesis of crystalline CaTiO3 from the mixture with CaO easier than from the CaO-rutile system. Grinding the CaO-anatase mixture for 2 hours or more enables us to produce very fine particles of about 20nm in the first order mean size. The amount of CaTiO3 in the ground product increases with improving its crystallinity as the grinding progresses. Many CaTiO3 crystal grains of about 5nm are formed in the mixture ground for 2 hours, and they grow up in the prolonged grinding. The grain size of the CaTiO3 crystals reaches about 20nm by about 5 hours of grinding, and the grain boundary and lattice fringe become clear as the grinding progresses.

The second material is 3CaO·Al2O3·6H2O (C3AH6), which has been synthesized mechanochemically from the mixture composed of calcium hydroxide (Ca(OH)2) and pseudo-boehmite ( -AlO(OH)) powders by room temperature grinding using a planetary ball mill. Use of -AlO(OH) sample with inferior crystallinity is more favorable for the mechanochemical synthesis rather than that with well crystalline one. The time required to form C3AH6 from the Ca(OH)2 - -AlO(OH) mixture is much longer than that from the Ca(OH)2-gibbsite (Al(OH)3) one. Adsorbed water from air during grinding plays a significant role in the formation of C3AH6 from the former mixture. After water addition to the Ca(OH)2 - -AlO(OH) mixtures ground for various times, excess hydrated calcium aluminates such as C2AH8, C3AH8-12 and C2A0.5H6.5 are formed in the starting and the short time ground mixtures, while a few amount of these compounds is formed in these hydrated mixtures after prolonged grinding. Formation of these excess hydrated compounds, which belong to layered structural materials, is enhanced in the presence of free Ca, Al compounds and water. A mixture of CaO and silica-gel (SiO2) was subjected to grinding using a planetary ball mill, followed by heating to investigate the temperature for synthesizing para-wollastonite (CaO·SiO2). The MC treatment of the mixture brings about amorphous aggregates with almost homogeneous chemical composition. 2-hours MC treatment enables us to synthesize para-wollastonite by heating for 2-hours at 1273K, which is significantly lower by about 130K as usual. Heating the 5-hour ground mixture at 923K gives us to form a precursor of wollastonite, leading to its easy crystallization at higher temperature than about 1273K. Thus, it is found that the MC treatment for the mixture before heating is quite effective for synthesizing para-wollastonite.

The suggestion from the IFPRI members has come to the contractor for the second year’s work in 1999 on MC interaction between organic and inorganic materials. This is the initiation to start the solid-state reaction between polyvinyl chloride (PVC, [CH2CHCl]n) and CaO and/or Ca(OH)2 powders. The mixture was subjected to grinding using a planetary ball mill under different conditions, to investigate their MC reactions. The grinding causes dehydrochlorinating reaction, forming CaOHCl and [CH=CH]x. The reactivity against PVC of CaO is superior to that of Ca(OH)2, but all the same, the reaction yield is advanced as the grinding progresses. Furthermore, the yield and rate of the reaction are improved with an increase in the molar ratio of (CaO/PVC) as well as the rotational speed of the mill. Impact energy of balls would be also an important operational parameter governing the MC reaction.

In the third year (2000), the contractor has made the work for estimating the yield of MC processes by the use of ball mill simulation work based on the Discrete Element Method (DEM). This work enables us to find out the optimum condition of the MC process as well as the scale-up role of a MC reactor (a mill). The present study has been composed of three examples:

1. MC treatment of EP dust, forming soluble vanadium (V) compound in water,
2. MC treatment of fluorescent powder, accelerating its structure change, and
3. dechlorination of polymers with halogen by its MC treatment with inorganic material such as CaO.

Regarding the first example, the yield of vanadium extracted by water leaching is well correlated with impact energy of balls in the mill calculated from the result obtained by. As for the second example, the dry grinding the EP dust enables us to form a water soluble vanadium compound. The well correlation between the V-yield and the impact energy of balls is obtained, suggesting that the ball impact energy plays a significant role to control the formation of vanadium compound. The third one is dechlorination of polymers such as PVC (poly-vinyl chloride), PVDF (poly-vinylidene fluoride) and PTFE (poly-tetra fluoro-ethane) by their MC treatment with inorganic material such as CaO. The report described only the dechlorination of PVC and its correlation with the impact energy of balls in the mill calculated from the result simulated. All the same, the impact energy of balls in a mill is a significant key to control MC effect and reaction. In such sense, the computer simulation regarding the ball motion during milling is a quite useful tool for determining the optimum operational parameters, mill design with scaling-up.

The main aim of the project carried out in the years-1997-2003 was to deliver complex and profound, experimental and theoretical, description of co- and counter-current spray drying process including determination of drying and degradation kinetics, determination of final product properties and elaboration of our own CFD code for reliable scaling up of spray drying process. One of the most significant outcomes of the project was designing and building a 9 m long, 0.5 m in diameter spray drying tower. The tower capacity enabled identification of the effect of various drying process parameters, like the effect of feed properties, feed rate and feed temperature, drying agent temperature and air flow rate on drying and degradation kinetics, particle residence time, particle morphology, etc. The column was equipped with a 72 kW heating system, waste air cooling system, dust collecting system and optical glass windows to perform measurements using laser techniques (LDA, PDA). Feed was delivered from a stainless steel tank (equipped with a cooling/heating steam/water jacket, 150 liter volume) to the nozzle by a progressive-cavity pump. Two steam generators were used to deliver heating agent to the jacket to maintain the required temperature of the drying material. Each pipe and the nozzle were equipped with a water heated jacket. The construction of the tower enabled taking samples at subsequent time intervals and making laboratory analysis of moisture content, size distribution, etc., as well as the quality index specific for a given product at a different distance from the atomizer. To determine the flowfield in the spray tower and structure of the spray in the cross-sectional area and along the length of the tower, the laser technique was employed. The tower walls contain numerous portholes to allow laser measurements at different vertical locations. A transport system of laser device comprising hoists and pulleys enables the laser unit to move easily between levels and carry out measurements at an arbitrary height of the column and in selected points in a given cross section. First three years of the project were devoted to development and improvement of experimental equipment to perform in situ measurements of drying process parameters using Particle Dynamic Analysis (PDA) to determine the structure of spray and microseparator to find drying agent and product temperatures and humidities. This is definitely, the most sophisticated spray drying system ever developed for research of this process. Microseparator technique, significantly modified during the project realization can be easily used by industry people to find accurate temperature of a drying agent undisturbed by the presence of product particles moving in the air. Cocurrent spray drying operation mode was subjected to exhaustive experimental analysis in first three years of project realization. All relationships between initial drying process parameters and behaviors of continuous and discrete phase were determined. The results obtained throw a new light on the mechanism of the process which takes place inside the spray tower during drying of various materials. Most of the results were presented for the first time in literature. One of the most important findings of this work, in our opinion, was lack of aerodynamic segregation of particles in the drying chamber. The analysis of the results shows that in each point along the spray axis there is practically an identical particle size distribution. Spray is mixed very well from the very beginning of the process. This finding is of great importance for understanding the phenomenon of simultaneous momentum, heat and mass transfer during spray drying. One of the challenges of the project was to find drying and degradation kinetics in spray drying process. Substantial degradation of the products was found for most of the trials. Experiments showed a rapid decrease of baker’s yeast activity in the vicinity of the atomizer, the finer atomization the highest degradation of the product. Spray drying of heat sensitive products requires careful selection and control of process parameters.

To obtain a full picture of the mechanism of spray drying process, extensive experimental trials were extended to counter-current spray drying process. Opposite to the co-current spray drying process, due to complex hydrodynamics of continuous phase, an aerodynamic segregation of particles (more bigger particles close to the column wall) was found. Analysis of the results confirms literature suggestions about strong couplings of momentum transfer between continuous and discrete phases which cannot be neglected in modeling of counter-current spray drying process. An increase of mean particle size with the distance from the nozzle caused by agglomeration process in recirculation zones in the column was observed. The results proved high sensitivity of counter-current spray drying process to initial drying and atomization parameters and a position of the nozzle in the dryer. Generally, in relation to practical applications, we can conclude that a performance of the counter-current spray drying process is stable in a narrow range of process and atomization parameters which makes such a system difficult to control. Summarizing, we could conclude that complexity of spray drying process is an outcome of three factors: parameters of a drying agent (temperature, flow rate), atomization parameters and particle residence time in the column. Drying kinetics of spray drying process comes from difficult to predict influence of the above mentioned parameters. One of the leading ideas in the project extension was to develop a CFD model of spray drying process to point out why the existing codes, in a certain range of process parameters and given geometries of the spray drying chamber fail to predict the spray drying process. A spray drying column developed at Lodz Technical University, was employed to collect a database to verify a CFD model of spray drying process. A comparison of experimental and theoretical results of CFD modeling enabled us to formulate the conclusions which might be important for industry people to scale up or check the performance of spray drying process under different operating conditions.

Every CFD model of spray drying process which can be found in the literature or developed individually enables a relatively correct determination of the continuous phase parameters (e.g. distributions of drying air temperature and humidity) regardless a number of simplified assumption in initial and boundary conditions assuming that heat losses to environment and effect of atomizing air were taken into account and proper model of flow turbulence was selected. However, to obtain reliable results of CFD simulations concerning also the discrete phase parameters, it is necessary to introduce to the model real initial particle size distributions and mass flow rates of the disperse phase and real evaporation kinetics. This data are often difficult to obtain in industrial conditions. One of the main successes of the project was to construct and test a device for determination of disperse drying kinetics on a laboratory scale. Our main idea was to elaborate a system where we could reach evaporation rates (and drying times) similar to those obtained in a spray drying column. A laboratory drying tunnel equipped with effective heating system and stabile weight measuring system has been designed and built. The tests on drying of maltodextrin solutions proved repeatability and correctness of this method. In the developed unit, we finally gained evaporation rates and drying times similar to the conditions obtained in a drying column. An effect of the initial moisture content on the critical moisture content was observed which is related to a decrease of the equilibrium vapor pressure over the solution and a decrease of the driving force of evaporation and drying rate of the process. Results of the experiments proved that the generalized drying curves determined in the lab scale could be used in scaling up of spray drying process if the critical moisture content of the material is known. Outcome of this part of the work offers to practitioners a cheap, fast and precise technique to determine realistic spray drying kinetics from small scale experiments.

The last part of the project which delivered a complementary picture of the mechanism of the process consisted of complex experimental studies on the effect of drying and atomization parameters on the properties of selected materials during spray drying process. Materials from two groups were chosen for investigations, skin-forming (maltodextrin) and agglomerate-like (detergent and cacao), that are most often subjected to spray drying in industrial conditions. In the study, 352 experimental tests of spray drying of maltodextrin, detergent and cacao powder as a function of temperatures and air flow rate, atomization conditions as well as temperature and dry matter content in the feed were carried out (in the last two years). We found that for all powders the mean particle diameter was smaller than the mean droplet diameter in the spray which was caused by shrinkage of the particles during moisture evaporation. We also proved that, depending on drying and atomization conditions, each of the tested materials could be cohesive, slightly cohesive or loose. We determined quantitative relationships and explained effect of the initial distribution of particle diameters and their morphology on product bulk density. Determination of these conditions in the frame of this work has an important practical meaning. The work delivers a profound and complementary picture of the mechanism of the process which takes place inside the spray tower during drying of various materials. The behavior of dispersed phase, drying and degradation kinetics, quantitative relations between initial parameters of drying and atomization and physical properties of spray-dried products are presented for the first time in literature. Determination of the local particle size and velocity distribution, detection of the uniformity of spray structure in the dryer, elaboration of quantitative relations describing drying and degradation kinetics are among the most important outcomes of the project.

Executive Summary:

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. In part I we demonstrate that true nanoparticle dispersions can be modeled and studied as colloidal dispersions. These results show that the continuum hypothesis of hydrodynamic interactions holds down to the nanometer scale. The research provides guidance on the formulation, rheological investigation, and modeling of nanoparticle dispersions. Part II completes a study on the effects of particle shape on colloid rheology. A model system of industrial relevance is characterized and studied. Rheology and neutron scattering under flow show that the mechanisms of reversible shear thickening in concentrated colloidal dispersions of ellipsoidal shaped particles are identical to that in colloidal dispersions of spherical particles. We show, however, that shear thickening appears at lower volume fractions in highly anisotropic particles and demonstrate a scaling that tracks the isotropic-nematic transition. Furthermore, the neutron scattering results demonstrate that shear thickening is not a consequence of flow-induced disorientation in these dispersions, dispelling that hypothesis and providing a quantitative means for predicting the onset of shear thickening in highly anisotropic dispersions. Part III provides a look forward for extending the research results obtained herein. It also includes results for clay dispersions that shows the universal nature of reversible shear thickening in spherical, prolate, and oblate particle dispersions.

The results of this body of work enable modeling and predicting that behavior of concentrated dispersions of colloidal and nano-sized particles that can be of assistance in formulating products and controlling processes involving dispersions.

This year's research on the behavior of concentrated suspensions has focused on the connection between viscous suspensions and granular media - from wet to dry - as it pertains to mixing. Here 'mixing' is quantified in its most primitive form - the diffusive motion of the particles. Diffusion is one of the most basic and elemental transport processes and is responsible for the molecular mixing of different chemical species. For a small sub-micron- or nano-sized colloidal particle, the diffusivity is given by the Stokes-Einstein formula relating the diffusivity to the thermal energy times the hydrodynamic mobility. The self-diffusivity decreases as the concentration of nanoparticles increases owing to the crowding effect of near neighbors. As the diffusing species increases in size from a nanoparticle to a several micron-sized colloidal particle, the stirring of the background fluid can give rise to another mechanism of transport - 'shear-induced' diffusion. Here, hydrodynamic interactions among particles promote mixing and the self-diffusivity now scales as the particle size squared times the shear rate. In this regime, the self-diffusivity is an increasing function of concentration since particle-particle 'collisions' are responsible for the diffusive motion. At still large particle size (millimeter or larger), the inertia of the particles becomes important, direct particle-particle collisions dominate, and the self-diffusivity now behaves like that in a dense gas with the diffusivity proportional to the mean-free path times the square root of the 'granular temperature', the latter of which is set by the stirring motion and the energy dissipated upon collision. As in a dense gas, the self-diffusivity now decreases with increasing particle concentration. The physical origin of these various behaviors, the dependence of the self-diffusivity on shear rate and particle concentration and their implications for mixing and particle distributions in inhomogeneous flows is discussed in this report.

Executive Summary

This report summarises the work done over six years in IFPRI project 37 Quantitative Analysis of Powder-Binder Agglomeration. This is a large body of work undertaken by ten researchers and IFPRI funds were substantially leveraged with funds from other sources. The main focus of the report is the analysis, in turn, of each of the three classes of granulation rate processes that dictate the product granule attributes:

1. Wetting, binder dispersion and nucleation
2. Consolidation and growth
3. Wet granule breakage

We now know the key formulation properties and process parameters that control the rate processes of (1) nucleation and wetting, and (2) consolidation and growth. For both these rate processes, regime maps have been developed and validated based on the controlling dimensionless groups: a and tp for wetting and nucleation, Stv, Stdef, and s for growth and consolidation. This quantitative understanding of granulation rate processes is now at the point where it can be directly used in scaling granulation processes (eg. keeping dimensionless spray flux constant to maintain nucleation conditions) and characterising formulations for their granulation behaviour (eg measuring the dynamic yield stress of a new formulation).

The report gives details of the development of characterisation tools, models and regime maps used to quantify the rate processes and experimental verification of models and regime maps. The impact of both process parameters and formulation properties on the granulation rate processes is studied in detail.

We are applying the same approach to quantify wet granule breakage, although this work will not be complete till the end of 2004. We report an approach to characterize the mode of failure of the granule matrix (brittle crack propagation or plastic deformation) for different formulations using an Instron Dynamite testing rig and detail the design and methodology for studying wet granule breakage in a breakage only granulator (BOG).

The report also gives summaries of related research in two areas, not directly funded by IFPRI, but very relevant to our multiscale approach to granulation modeling:

• Using positron emission particle tracking (PEPT) to characterize powder flow fields in granulators and using this information in conjunction with rate process models to predict granulator performance;
• Design and modeling of regime separated granulators for significant improvement in the control of granule attributes. Finally, a discussion of research and development gaps and opportunities in the field of granulation is presented.

Executive Summary

The primary aim of this project is to provide fundamental information regarding the manner in which agglomerates of fine particles disperse in response to the application of hydrodynamic shear. A major emphasis of the research supported under this IFPRI grant involves investigation of the role of certain time-dependent (dynamic) phenomena in governing dispersion associated with non-steady shear flows or fluid wetting phenomena. In addition, we are interested in assessing the role of binders in governing dispersion behavior. We also developed predictive analytical models for the various modes and kinetics of the dispersion process. Three principal experimental tools have been used throughout this work. One allows the observation of dispersion in steady simple-shear flows of controllable intensity. A second other enables the application of a time-varying shear stress with controlled frequency and amplitude. The third technique enables sensing of the micromechanical behavior of particle clusters and/or the forces associated with the deformation of liquid bridges between individual pairs of particles. This report details progress in five parallel and related thrust areas:

1. The relationships between agglomerate characteristics, flow conditions, and dispersion phenomena (mode and kinetics) for “dry” agglomerates. This topic includes the development and validation of a new model for predicting dispersion kinetics.
2. The identification of a new mode of dispersion (adhesive failure) and an analysis of the conditions under which it occurs.
3. The importance of flow dynamics (time-varying flows) on the dispersion process. This topic also includes the development of a mathematical analytical approach for analyzing the influence of flow dynamics on dispersion tendencies.
4. Characterization and analysis of the micromechanics of particle clusters, including those with interstitial fluids and/or liquid bridges. This topic also leads to a modeling approach useful for the understanding and prediction of dispersion mode.
5. Development of a detailed, fundamental model that provides precise information on the dispersion of clusters ranging in size down to the nanometer scale.

Executive Summary

The main aim of this project was to develop the necessary understanding, databases, guidelines and models for the purpose of predicting accurate optimal operating conditions for two modes of dense-phase conveying: low-velocity slug-flow (LVSF) of granules and fluidised dense-phase (FDP) of powders. During the first 3 years of the project, top priority was given initially to LVSF, although some progress also was made with the FDP section of work. The second 3-year period of the project allowed a substantial amount of work to be completed in the FDP section, as well as some interesting investigations emanating from the LVSF work.

Several difficulties were encountered during the course of the project (e.g. unexpected results and phenomena) and these delayed progress in various areas. In some cases, it was not possible to complete certain planned tasks (e.g. testing different pipe wall materials for the LVSF section and also a wide range of bulk solids for the LVSF and FDP sections, further testing and modelling for the FDP section). In other cases, it was necessary to pursue new research activities (e.g. rotary valve air leakage, new pipe wall friction and stress transmission testers, comparisons with other FDP models). However, in terms of achieving the overall goals, there is no doubt that the 6-year project was successful, as measured by the following achievements:

• improved understanding of both LVSF and FDP;
• development of new and comprehensive databases for both LVSF and FDP;
• development of guidelines and new fundamental models for the prediction of LVSF performance (e.g. horizontal pipeline pressure drop, unstable boundaries);
• evaluation of and development of guidelines for FDP performance (e.g. minimum transport conditions, empirical modelling of powder flow frictional properties, scaleup accuracy of new and existing models);
• as well as some new and exciting (additional) developments (e.g. new particle/bulk property testers; novel comparisons of blow tank and high-pressure rotary valve feeder performance for LVSF; new on-line instrumentation to monitor LVSF performance; evaluating the scale-up stability and accuracy of existing FDP models).

Background and Project Overview

Atomization is a chemical engineering unit operation, which disintegrates a continuous liquid into a dispersed system of drops within a spray. Common atomizers for disintegration of liquids are producing droplet spectra to be characterized by:

• droplet size distribution,
• droplet size/velocity correlation and
• local and overall concentration and mass flux distribution.

These major describing parameters of a droplet spray are mainly influenced by relevant parameters such as:

• the atomizer design and working principle,
• fluid material properties and
• mass flow rates.

Within an existing atomization process, manipulation of the spray properties and droplet spectra often only is possible by means of changing the energy input into the atomizer, which simultaneously alters several of the relevant spray parameters with a minimum degree of control. This research project aims to analyse the influence of solid particles on the droplet characteristics in the liquid atomization with suspended particles and to develop atomizer strategies and design technical equipment in order to produce controlled droplet characteristics within a wide range of applications. Various model suspensions based on water, water/glycerol and/or water/Non-Newtonian fluid mixtures with particles will be atomized by means of twin-fluid atomizer, rotary atomizer and pressure jet nozzle. The influence of particle characteristics and loading of particles on the liquid jet and liquid film disintegration are analysed.

Executive Summary

The basic concept in the project research to address the electrostatic charging of powder was that it is essentially important to study the impact (contact) charging of a single particle due to a single collision. To realize the concept, we proposed two approaches. One was a development of our original ‘impact charging experiments method’ (approach 1). The other was a new technique involving the measurement of the electrostatic adhesive force curve (approach 2).

Approach 1:

• A new equipment was designed and successfully constructed.
• The sensitivity of the charge measurement was improved by 1000 times, which allowed impact charging experiments to be made on particles from 50 to 500 µm.
• Although the actual data were scattered, almost all of them fell in the region that can be explained by the ‘charge relaxation model,’ with an assumption of localized initial charge.
• Surface conductive particles lost their all the initial charge at the impact onto a metal target: this was the same as found with 3 mm metal particles in the previous work.
• Experiments with different kinds of metal targets with different work functions did not show any significant difference; this is not expected from the conventional model regardless of the charge relaxation path.
• From the results, it was shown that the charge relaxation model works at this particle size range of hundred microns as well as the case of bigger particle with 1-3 mm in diameter in the previous work.

Approach 2:

• The force curve was successfully measured with atomic force microscopy (AFM).
• By detailed analysis of the ‘force curve’ measurement, rather than only looking at the maximum adhesive force, the electrostatic interaction was successfully observed by separating other interactions such as liquid bridge and intermolecular force.
• A theory was developed to evaluate the force curve based on an image force method using an approximation of disk-to-disk interaction.
• The good agreement between the observed force curve and theory showed that the force curve observed can be surely attributed to the electrostatic interaction, and that the amount of charge on the particle and the radius of the charged (contact) area can be estimated from the analysis.
• The order of magnitude of the measured charge density was 10-2 C/m2, which is much greater than that obtained with impact charging experiments (10-4 C/m2). The force curve measurement with AFM can catch the net amount of the charge generated before charge relaxation due to gas discharge or charge relaxation takes place.
• In the experiment using 8 kind of metal targets, the net charge generated showed fairly good agreement with the conventional simple condenser model, but the data dispersed.
• Behind the apparent random data scatter, a strong relationship was found between the charge density and contact area. The mechanism of the relationship is not know at this moment, however the finding is a good starting point for future work.