Powder Testing Guide: Methods of measuring the physical properties of bulk powders

POWDER TESTING GUIDE Methods of Measuring the Physical Properties of Bulk Powders L. SVAROVSKY School of Studies in Pow...

Author: L. Svarovskly

157 downloads 580 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

POWDER TESTING GUIDE Methods of Measuring the Physical Properties of Bulk Powders

L. SVAROVSKY School of Studies in Powder Technology, University of Bradford, UK

Published on behalf of the

BRITISH MATERIALS HANDLING BOARD by

ELSEVIER APPLIED SCIENCE LONDON and NEW YORK

FI.SI'VII;,R APPLIED SCIENCE PUBLISHERS LTD ( 'I< 111'11

I III II St' • Linton Road, Barking, Essex IG 11 8JU, England

Sole Distributor in the USA and Canada ELSEVIER SCIENCE PUBLISHING CO., INC. 52 Vanderbilt Avenue, New York, NY 10017, USA WITH 1 TABLE AND 39 ILLUSTRATIONS

© BRITISH MATERIALS

HANDLING BOARD 1987

British Library CalalllJ.:llillJ.: ill Publication Data Powder testin!!, guide: mcthods of measuring the physical propertics of bulk powders. I. Powders I. Svarovsky, Ladislav II, British Materials Ilandling Board h20'A3 TA418.78 Library of Congress Cataloging-in-Publication Data Svarovsky, Ladislav. Powder testing guide. Bibliography: p. Includes index. 1. Powders-Testing. Title. TA418.78.S83 1987

r.

British Materials Handling Board.

620'.43

II.

87-15722

ISBN 1-85166-137-9 No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Special regulations for l'eaders in the USA This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside the USA, should be referred to the publisher. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Printed in Great Britain by Page Bros (Norwich) Ltd

Foreword

Standard methods of measurement of physical properties of powders in relation to bulk handling and processing have been adopted for specific materials and have sometimes been incorporated into a BS or other Standard. However, general standardised methods have yet to be established and the BMHB believes that much wider applications could be made of proven techniques which, while being well established in specific industries, have yet to be generally exploited. The aim of this Guide is to review, detail and recommend preferred test methods, to outline their significance and to identify test methods on powder characteristics which require further research, development or evaluation. The accent is on simple and inexpensive techniques. The Guide draws on existing Standards, Trade and Research literature and on expert knowledge of selected designers and manufacturers of powder handling plant. It is anticipated that the Guide will prove valuable in disseminating knowledge on simple powder testing techniques and assist in establishing better communication between equipment users and suppliers. The Guide was prepared by Dr L. Svarovsky, University of Bradford, with the help of a British Materials Handling Board Steering Committee of the following membership: Dr P. C. Knight, Unilever Research Dr D. Geldart, University of Bradford Mr R. E. Pace, Simon Engineering v

Author's Acknowledgements

I am very grateful to all of the industrialists and academics I have talked to during the preparation of this Guide. I must acknowledge with particular thanks the following people who have made available to me their own notes and written material about the details of test methods quoted in this Guide: Dr J. C. Williams and Dr D. Geldart of Bradford University, Mr L. Bates of Ajax Equipment (Bolton) Ltd and Mr M. Walton of Chr. Michelsen Institute. An acknowledgement also has to be given to the authors of the Bulk Solids Physical Property Test Guide published by the BMHB as much of the background reference was drawn from this publication.

vii

Scope

The Guide confines itself to dry solids (as opposed to wet cakes) but includes the effects of air humidity. It deals mostly with powders, typically finer than about 3mm, and it excludes detailed description of electrical and thermal properties and explosion/fire hazard testing. The reader is referred to Ref. 82 for these. The emphasis is on bulk or "technological" properties of powders and the primary properties like particle size, shape and distribution are treated only as background. The problems and importance of sampling are included but merely for guidance rather than in technical detail.

Contents

Author's Acknowledgements.

vii

Scope

ix

Introduction

1

1 Sampling

3

2 Properties Dependent on Single Particle Characteristics 2.1 2.2 2.3 2.4

Selection of Relevant Characteristic Particle Size. Description of Particle Size Distribution and Mean Size Particle Shape Particle Density . 2.4.1 Measurement of Particle Density 2.4.1.1 Liquid Pyknometry . 2.4.1.2 Air Pyknometry 2.4.2 Measurement of Effective (Aerodynamic) Particle

W

~m~

2.4.2.1 Caking End Point Method 2.4.2.2 Bed Voidage Method 2.4.2.3 Bed Pressure Drop Method 2.4.2.4 Sand Displacement Method 2.5 Surface Area of Powders 2.6 Moisture Content

3 Categorisation of Powders According to Behaviour in Handling . 3.1

11

12 13 14 16 17 17 17

Classification of Powders in the De-Aerated State

xi

20 21 22 23 23 29 35 35

xii

CONTENTS

3.1.1 Classification on the Basis of Shear Cell Testing 3.1.2 Classification on the Basis of Tackiness 3.2 Classification of Powders in the Aerated State 3.3 Classification of Powder Handling Properties in Pneumatic Conveying

36 37 38 40

4 Non-aerated Flow and Handling Properties 4.1 Definition of Failure Properties 4.2 Angle of Wall Friction. 4.3 Angle of Internal Friction 4.3.1 Shear Cells 4.3.2 Biaxial and Triaxial Shear Testers 4.3.3 Direct Method (Grooved Plate) 4.4 Failure Function . 4.4.1 Indirect Methods 4.4.2 Direct Methods 4.4.2.1 Uniaxial Compression-William's Method 4.4.2.2 Compression Tackiness Tester. 4.4.2.3 Large-Scale Uniaxial Test 4.5 Tensile Strength . 4.5.1 Split Cell Testers 4.5.2 Lifting-Lid Testers 4.6 Cohesion 4.7 Angle of Repose and Other Handling Angles 4.7.1 Angle of Repose of a Heap 4.7.2 Drained Angle of Repose. 4.7.3 Angle of Slide . 4.7.4 Conveying Angle 4.7.5 Angle of Sliding 4.7.6 Angle of Spatula 4.8 Flowability and Flowrate Tests

41 41 47 48 49 52 53 54 55 55 55 58 61 62 63

5 Packing Properties, Bulk Densities . 5.1 Porosity of a Packed Bed, Void Ratio.

79 79 81 81 83 84 89

5.5.1

Measurement of Porosity .

5.2 Bulk Density of a Powder 5.2.1 Aerated Bulk Density 5.2.2 Poured Bulk Density 5.2.3 Tap Density 5.2.4 Hausner Ratio .

66

67 71 73 74 75 75 76 76 77

92

CONTENTS

5.2.5 Compressibility (from Bulk Densities) 5.2.6 Compacted Bulk Density . 5.3 Compaction Tests

6 Grinding and Strength Properties 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8

Single Particle Strength Hardness Grindability Tests Impact Tests Compression Tests Vibration Tests . Abrasion Tests FriabilityI Attrition Tests

7 Aerated Flow and Handling Properties 7.1

7.2 7.3 7.4 . 7.5 7.6

Fluidization Tests 7.1.1 Equipment for Fluidization Tests When Using Fine Powders . 7.1.1.1 Column Size 7.1.1.2 Distributors 7.1.1.3 Leak Testing 7.1.1.4 Pressure Drop Measurement 7.1.1.5 Air Supply 7.1.2 Measurement of Minimum Fluidization Velocity 7.1.3 Minimum Bubbling Velocity 7.1.4 Bed Expansion. 7.1.5 Bed Collapse, De-Aeration Rate Simple De-Aeration Tests Permeability Dustiness Floodability. Apparent Viscosity Tests

xiii

93 93 93 97

98 98 99 103 104 104 104 106 111 111 112 112 112 113 114 114 115 116 117 118 120 121 122 126 127

8 Conclusions, Future Work

131

9 References.

133

Index.

139

INTRODUCTION

In the process of compiling this Guide, it became clear that powder testing can be put into three categories as follows:Category A where standardised tests already exist and the interpretation of data is well established and accepted; the only work needed is in publicising their existence and use. Category B where test methods exist but there is disagreement as to the significance of the measured values in equipment design and scale-up; validation testwork is clearly needed here. Category C where no standardised test procedures are yet available for a particular powder property and these need to be developed experimentally. The Guide is not written in the order of the above categories but they are assigned at the end of the discussion of each test method. The bulk powder tests are reviewed in the order of the following groups of powder properties to which they relate: Properties dependent on single particle characteristics Properties of non-aerated powders Packing properties, bulk densities Grinding and strength properties Properties of aerated powders Prior to embarking on the review, as a general observation, it is interesting to note that there is an imbalance in the area of powder testing in favour of very detailed particle size and shape 1

2

POWDER TESTING GUIDE

measurements using very sophisticated and expensive laboratory instrumentation, without yet having the fundamental correlations to translate the measured data into secondary behaviour. With the notable exceptions of research organizations or large companies with specialised goals and applications, such intense preoccupation with very detailed physical data is often made at the expense of testing for other important powder properties. The number and quality of the books available on particle size measurement, in contrast to those on other powder testing methods, is a good indication of this imbalance; it is hoped that this Guide will help to redress the balance. Another point to be made here concerns the effects of gas properties. Generally, the effect of the interstitial and surrounding gas on the mechanics of dry solids handling is neglected. The effects of gas properties may be two-fold: 1. An aerodynamic interaction between the gas and the solids, mainly controlled by the dynamic viscosity of the gas and the elasticity of the packed solids. 2. A physical-chemical interaction through adsorption of the gas on the solid surface, which affects the resistance to breakage, attrition and abrasion. Thus the properties of the suspending gas have to be taken into account, or controlled, not just in the obvious applications where gas clearly plays a part like fluidization or pneumatic conveying but also in the not -so-obvious ones like grinding and discharge of powders from hoppers. Moisture content in the gas is known to affect most solids handling properties to the extent that, if powders cannot be kept reasonably well sealed, air-conditioning is necessary if meaningful data are required from powder tests.

1 Sampling

It is often said that any test on a sample of powder can be only

as good as the sampling technique used for collecting the sample. As most laboratory tests use only a small sample, this has to be taken from a production stream or from an existing, stored material and it has to be representative of the whole. Unlike fluids, the properties of powders are susceptible to change under applied load, they may consolidate with time, and attrition and segregation occur in transfer. In particular, the facts that powders have a size distribution which affects so many of the powder properties so much and that segregation of stratification by size is so common, make representative sampling absolutely critical for the success and relevance of any subsequent testing. Sampling is therefore such an important element of powder handling that it demands careful scientific design and operation of the sampling systems. The purpose is to collect a manageable mass of material (= sample) which is representative of the total mass of powder from which it was taken. This is achieved by taking many small samples from all parts of the total which, when combined, will represent the total with an acceptable degree of accuracy. This means that all particles in the total must have the same probability of being included in the final sample. All parts of the total have to be equally accessible. To satisfy the above requirements, the following basic "golden" rules of sampling should be followed whenever practicable.

4


1. Sampling should be made preferably from a moving stream (this applies to both powders and suspensions) but powder on a stopped belt can be sampled . 2. A sample of the whole of the stream should be taken for many (equally-spaced) periods of time rather than part of the stream for the whole of the time . It is very likely that the re-combined, primary sample taken

from the whole is going to be too large for most powder tests and it, therefore, needs to be sub-divided into secondary or even tertiary sub-samples. This sub-division may be built into the primary sampling system or it may be achieved with a separate sample divider. AlIeni reviewed and tested most methods available for sample splitting and found the one based on the spinning riffler to be the best.

FIG .

1. A view of a spinning riffter.

The principle of the spinning riffler is shown in Fig. 1 and it embodies both golden rules of sampling. The sample is slowly conveyed by a vibratory feeder from the feed hopper onto a

SAMPLING

5

rotating carousel where it is divided into many container ports via a machined rotary head. The sub-samples are collected in these; not all of the container ports may have the containers in (depending on how many sub-samples are required) and where they have not, the powder falls into a bucket underneath. Feed rate is controlled by varying the gap under the hopper and varying the electro-magnetic vibration of the feeder. Many different commercial instruments based on this principle are available, and they may be built to divide as little as 25 ml or as much as 40 litres of powder or more. Whilst it is possible to design a range of small sample dividers for general use, the sampling systems for the collection of primary samples from large-scale processes have to be designed specifically for a particular material and application; a short review of such systems is given below. MECHANICAL SAMPLING Mechanical, as opposed to manual, sampling is usually preferred because it collects samples with better overall precision, at accurate time intervals a!1d can handle the samples and the whole operation automatically. The primary samples are usually collected by a primary cutter which cuts the full cross-section of the stream in a way similar to that of each compartment in a spinning riffler (Fig. 1) cutting across the falling stream of powder. The design of such cutters is subject to guidelines laid down in several recent publications. 2,3,4,5,6 Thus, for example2 :

* The minimum mass of the primary increments depends on the mode of operation (constant mass or constant time between the incremen ts ). * The minimum aperture is related to the nominal top size D of the material being sampled (= 3 x D, refer3 to BS 1017 for the precise definition of D). * The cutter lips should be normal to the mean trajectory of the stream and of such shape that each part of the lip is in the stream

6


for the same period of time (i.e. linear cutters should have parallel lips whilst radial cutters should have radial lips). * The cutter velocity should be less than 0.6 rn/s if the minimum cutter aperture of 3D is used but can be increased for larger apertures according to an empirical formula 2 :

where

is maximum velocity of the cutter (m/s) W is the actual width of the cutter aperture WI is three times the nominal top size D of the material being sampled, i.e. WI = 3D. Dc

It should be pointed out here that the above quoted limits of minimum cutter width of 3D and the maximum cutter velocity of 0.6 rn/s have not yet been accepted universally but Gt has published some experimental work to support this recommendation. The US standard ASTM D 2234 recommends the limit of 18 inls (0.457 m/s) whilst the Australian Standard AS2646 allows cutter velocities up to 1.5 mls except for secondary (and subsequent) sampling stages when the limit is 0.6 m/s; both standards referring to sampling of coal. It is absolutely essential, of course, that the cutter velocity is constant during sampling in order to avoid bias; the maximum permissible deviation in the velocity is usually quoted as 5%.

Types of Cutters The cutters in use in mechanical sampling are divided into diverter types and bucket types. Both types are used to cut a stream falling due to gravity off the end of a conveyor belt or from the discharge end of a pneumatic conveying pipe or a chute. The diverter cutters divert the stream increment clear of the main stream and, providing they are properly designed, they do not allow accumulation or sticking up of the powder anywhere inside; they require considerable head-room, however, and can only deposit the increment below the point of sampling and not very far from it laterally.

SAMPLING

7

The bucket-type cutters have the advantage of being able to collect and transport the sample laterally, without the loss of headroom; they collect and hold the sample, however, and thus allow material build-up within the bucket if the powder is a little sticky. Another disadvantage is that, whilst the mass of the diverter-type cutter remains the same during its traverse across the falling stream, the mass of the bucket cutter increases rapidly and the drive systems must be powerful enough to maintain its speed. As to the different designs available in each of the two groups, the reader is referred to two excellent recent publications by Plowman 2 and Merks5 • One design worthy of a special mention is the cross-belt type cutter (or rotating hammer sampler) which swings in plane perpendicular to the movement of a conveyor belt and scoops a well-cut sample off the belt. Unlike the other, linear cutters, this one is not limited in its speed but it should traverse the bed of powder on the belt in the shortest possible time. As was mentioned earlier, the primary sampling system can be operated either in intervals of constant time or constant mass. The constant mass option makes the design and operation of the secondary, subdivision system simpler. It requires a continuous weighing system, like a belt scale, installed near the primary cutter, preferably before it. This monitors the mass flux of the solids conveyed and adjusts the speed of the primary cutter before each cut; this generates a primary increment of constant mass, thus preventing collection of excessive amounts which would overload the secondary system and yet always more than the minimum amount specified for the top size if the handling rate is low. The secondary sample dividers are used to reduce the size of the primary sample; they can be classified5 as intermittent and reciprocating cutters, and continuous and rotational dividers. The rotational dividers are usually considered more suitable than linear cutters. Most standards define the minimum number of cuts that the secondary divider has to take from each primary increment: ASTM D2234, for example, requires six secondary increments to

8


be collected from each primary increment (for coal). This indirectly controls the division ratios from 1/16 to 1/37. This may not be enough to reduce the sample to an acceptable size (the final system sample should not exceed 20 kg) and tertiary or quartenary dividers may be necessary in large scale systems. Merks 5 recommends that ratios in excess of 1/50 in those stages are not acceptable because they would cause the variance of division to be excessively high. One aspect of mechanical sampling that is particularly relevant to this guide is bias testing 6 • These are procedures agreed between buyers and sellers in bulk solids handling designed to evaluate the performance of mechanical sampling systems. Such evaluation is based on a statistical comparison of paired measurements in system samples (i.e. samples taken by the mechanical sampling system) and reference samples taken manually (usually from a stopped belt, see below). The reference increments are usually collected in pairs at a spacing of 30 m or less (if from a belt) in such a manner that one system increment is interposed between a set of two reference increments. The series of paired measurements is then evaluated statistically (usually by Student's T-test) to determine the correlation between the analyses (particle size distribution or some other parameter) of the reference increments and the system increments, to determine any bias (positive or negative) and compare it with a previously agreed maximum permissible bias.

MANUAL SAMPLING Manual sampling is usually performed in low capacity handling and when the top size of the material is low. It can be done from a falling stream, from a stopped belt or from a stationary pile or hopper. The first option, from a falling stream, is usually used at a transfer point between conveyor belts, from under a discharging hopper or from the end of a pneumatic conveying system. Open ended scoops or shovels are unsuitable for manual sampling

SAMPLING

9

because they allow coarser particles to roll out of the sample and thus bias the sample towards the finer fractions. A ladle is the only recommended implemene for such sampling because it does not allow this rolling off to happen, unless overfilled. The falling stream is traversed with the ladle to collect the sample in much the same way as the mechanical cutters are designed to do and if the whole of the stream cannot be sampled in one pass, it is divided into several equal areas and the sample is collected from those incrementally. Stopped belt sampling is considered the best manual sampling method. It is often taken as a reference method and others are compared against it. It follows the two golden rules of sampling in that it samples from a moving stream (with the movement momentarily stopped for taking the sample) and it takes the whole of the stream many times. It is done with a suitable profiled sampling frame which is inserted through the material on a stopped belt conveyor until it comes in contact with the conveyor belt over its full width. The material within the frame is then collected from the belt and it represents one increment in the manual sampling series. Stopped belt sampling is frequently used in mechanical sampling system bias testing as mentioned above. If the total of the powder cannot be moved and the sampling has to be done from a stationary hopper or container, the samples have to be withdrawn from small spaces of equal volume within the total. This is very difficult to achieve because not all of the material is equally accessible. For sampling from relatively small volumes, sampling "thieves" are available which allow opening and closing of a small sample holder at the end of a pole inserted into the powder. A British Standard for testing cement? specifies a sampling ladle and a sampling tube for a similar purpose. Even this method, however, is not entirely satisfactory and sampling from stored powders is generally to be avoided. Sampling of coal and powder in general from trucks, for example, sometimes has no alternative and some guidance to the sampling technique exists2 • The· minimum width of the scoop is 2.5 times the largest size of coal to be sampled. Trucks are divided longitudinally into three equal strips and increments are taken

10


equally spaced within each truck and over the number of trucks sampled. The increments are taken by boring holes 300 mm deep in the coal and taking the sample from the bottom of such holes. In order to avoid the surrounding material falling into the hole, an open-bottomed frame may be used which is forced into the coal mass and the coal is scooped out of the bottom. Two other alternatives exist in sampling of coal from trucks: sampling by probe, when full depth insertion and withdrawal are used, and sampling by suction sampler. The latter is a sheathed coal-drill whereby the drillings are conveyed by suction to a sample container. This method is applicable only to coarse coal and should be agreed by the interested parties. Sampling from deep hoppers and open stockpiles is the most difficult and inaccurate of all sampling operations and it should be avoided whenever possible.

2 Properties Dependent on Single Particle Characteristics

The gap between theory and practice in powder storage and handling is still wide. It is very rare in any branch of powder technology that a designer can choose and size equipment from first principles, using the primary properties of the particulate system and of the suspending gas. The reason for this is that, due to the complexity of the processes involving particulate systems, there are no reliable quantitative relationships between primary properties like particle size and distribution, particle shape etc. and secondary properties such as failure properties or bulk density. Whatever the principle or equipment to be used for powder handling, reliable sizing or even selection of the equipment can only be done on the basis of small scale tests on the actual powder, either in a laboratory or in a pilot plant (unless large scale data with a similar powder exist). This is where the theory plays a very important part because it provides us with tools for the necessary scale-up procedures. It is not within the scope of this report to give a full account of the basic tests and scale-up theory but there are some fundamental concepts in particle characterization that have to be stated here. The characterization of primary particle properties in a particulate system and their correlation with the secondary bulk properties of the system is a problem common to all branches of particle technology. As pointed out above, it is still necessary in powder handling to measure the secondary properties directly and it can be argued, therefore, that particle characterization can be 11

12


by-passed. The properties of the carrier gas (like viscosity, density and moisture content) complicate the situation further and it is therefore not surprising that in practice this by-passing is quite common. However, primary particle properties still have at least a qualitative value, as .a selection guide, and they should be measured wherever possible, otherwise we shall never get nearer to finding out how things really work and to making theory work. The following short account of particle properties and how they are characterized only points out the basic philosophy and gives an over-view. More detailed study and information can be obtained from the literature 8,9.

2.1 SELECTION OF RELEVANT CHARACTERISTIC PARTICLE SIZE An irregular particle can be described by a number of sizes. There are three groups of definitions: the equivalent sphere diameters, the equivalent circle diameters and the statistical diameters. In the first group are the diameters of a sphere which would have the same property as the particle itself (e.g. the same volume, the same settling velocity, etc.); in the second group are the diameters of a circle that would have the same property as the projected outline of the particles (e.g. projected area or perimeter). The third group of sizes are obtained when a linear dimension is measured (usually by microscopy) parallel to a fixed direction. There is a wide variety of methods for particle size measurement which measure different types of particle size. When selecting a method, it is best to take one that measures the type of size which is most relevant to the property or the process which is under study. Thus, for example, in powder elutriation, pneumatic conveying or gas cleaning, it is most relevant to use one of the sedimentation methods which measure the Stokes' diameter, i.e. the diameter of a sphere of the same density as the particle itself, which would fall in the gas at the same velocity as the real particle (assuming Stokes' law). In flow through packed or fluidized beds, on the other hand, it is the surface-volume diameter (or diameter

PROPERTIES DEPENDENT ON SINGLE PARTICLE

13

of a sphere having the same surface to volume ratio as the particle) which is most relevant to the aerodynamic processes.

2.2 DESCRIPTION OF PARTICLE SIZE DISTRIBUTION AND MEAN SIZE Very few, if any, practical particulate systems are mono-sized. Most show a distribution of sizes and, depending on the quantity measured, the distribution can be by number, surface or mass. Conversion from one type of distribution to another is theoretically possible but it assumes a constant shape factor throughout the distribution which often is not true and such conversion is in error. The conversions are therefore to be avoided whenever possible by choosing a measurement method which measures the desired type of distribution directly. Except for a few specialized applications like rating of filter media, the most relevant types in powder handling are usually the mass or the surface distributions. If a population of particles is to be represented by a single number, there are many different measures of central tendency or "mean" sizes. Those include the median, the mode and many different means: arithmetic, geometric, quadratic, cubic, bi-quadratic, harmonic (ref. 1) to name just a few. As to which is to be chosen to represent the population, once again this depends on what property is of importance: the real system is in effect to be represented by an artificial mono-sized system of particle size equal to the mean. Thus, for example, in precipitation of fine particles due to turbulence or in total recovery predictions in gas cleaning, a simple analysis may be used to show that the most relevant mean size is the arithmetic mean of the mass distribution (this is the same as the bi-quadratic mean of the number distribution). In flow through packed beds (relevant to powder aeration or de-aeration), it is the arithmetic mean of the surface distribution, which is identical to the harmonic mean of the mass distri bu tion.

14


2.3 PARTICLE SHAPE It is well known that particle shape affects many secondary prop-

erties relevant to powder handling such as the bulk density, failure properties or particle-gas interaction. For non-spherical particles, the results obtained with different methods of particle size measurement are, in general, not comparable. From the point of view of powder handling, flaky or stringy particles like wood shavings, mica or asbestos fibres are known to be difficult because they interlock and form obstructions to flow. A number of methods have been proposed for particle shape analysis; these include verbal description, various shape coefficients and shape factors, curvature signatures, moment invariants, solid shape descriptors, the octal chain code and mathematical functions like Fourier series expansion or fractal dimensions. As in particle size analysis, here one can also detect intense preoccupation with very detailed and accurate description of particle shape, and yet efforts to relate the shape-describing parameters to powder bulk behaviour are relatively scarce.1O It is also worth noting that, in many cases, particle size and shape do not exist independently, and shape can be size-dependent. This presents difficulties: for example, when conversions from one type of size distribution to another are necessary large errors are introduced because such conversions assume a constant shape factor throughout the whole size range. Some instruments even perform such conversions automatically and the user often does not realise the existence and the consequence of this assumption. As with the different definitions of particle size, the choice is made of a shape factor most relevant in the application in question. The following is a list of the definitions of the most frequently used, simple shape factors. SPHERICITY is the ratio of the surface area of a sphere having the same volume as the particle, to the actual particle surface area; the reciprocal is known as the coefficient of rugosity or angularity. It can be shown that sphericity is also equal to the ratio of the surface-volume diameter to the equivalent volume diameter; this makes sphericity a useful conversion factor between


15

the equivalent volume diameter as measured by instruments like the Coulter Counter and the surface volume diameter required in applications like fluidization, filtration or flow through packed beds. The range of sphericities for most angular solids is from 0.64 to 1, with near-spherical particles like round sand having values, of course, close to 1. Platelets like mica, on the other hand, may have a sphericity as low as 0.28. Tables exist8 of sphericities for regularly-shaped particles and one useful way of estimating sphericity of irregular particles is to compare the images under the microscope with the regular shapes for which spericities have been calculated. CIRCULARITY is the ratio of the perimeter of a circle having the same area as the projected area of the particle to the actual particle perimeter. This is clearly only a two-dimensional representation of a particle shape and as such can be evaluated by microscopy, preferably linked to an image analyser. This description of particle shape is not directly relevant to bulk properties of powders but could be useful in some applications. SURFACE-SHAPE COEFFICIENT is the coefficient of proportionality relating the surface area of the particle with the square of its measured diameter, the latter being one of the many possible definitions of particle equivalent diameter (this has to be defined when quoting values). This description of particle shape is useful in applications when particle surface is important. VOLUME-SHAPE COEFFICIENT is the coefficient of proportionality relating the volume of the particle with the cube of its measured diameter; the same qualifications apply here as in the case of the surface-shape coefficient but this time related to particle volume. SURFACE-VOLUME SHAPE COEFFICIENT is the ratio of surface to volume shape coefficients, combining together the features of the two. Apart from the basic definitions above, there are other shape factors which are defined as ratios of two different types of equivalent particle size. These are obtained by comparing particle size distribution measured by two different methods; the shape factor is the multiplier which would bring the results into coinci-

16


dence. Thus, for example, there is a shape factor relating the Stokes' diameter to the sieve diameter or to the equivalent volume diameter as measured by Coulter Counter or similar. An attempt to compare the sieve size with the equivalent volume and the surface-volume diameter for crushed quartz, as described by Abrahamsen and Geldart ll recently, also falls into this category.

2.4 PARTICLE DENSITY Particle density is the total mass of the particle divided by its total volume. Depending on how this total volume is defined (or measured), we can have the following densities (starting with the largest value): TRUE PARTICLE DENSITY when the volume measured excludes both open and closed pores. This is the density of the solid material of which the particle is made; for pure chemical substances, organic or inorganic, this is the density quoted in reference books of physical/chemical data. APPARENT PARTICLE DENSITY is when the volume measured includes closed pores or bubbles of gas within the particle. This density is measured by gas or liquid displacement methods like the liquid or air pyknometers (see below). EFFECTIVE (OR AERODYNAMIC) PARTICLE DENSITY is when the measured volume includes both the closed and the open pores. This volume is within an aerodynamic envelope as "seen" by the gas flowing past the particle; the value of density measured is therefore a weighted average of the solid and immobilised gas (or liquid) densities present within the envelope volume. The effective density is clearly of primary importance in applications involving flow round particles like in fluidization, sedimentation or flow through packed beds. Any of the three particle densities defined above should not be confused with bulk density of materials which includes the voids between the particles in the volume measured; bulk density of powders is dealt with in another section of this guide (section 5.2).


17

The different values of particle density can also be expressed in a dimensionless form, as "relative density" (or specific gravity), which is simply the ratio of the density of the particle to the density of water. 2.4.1 Measurement of Particle Density The apparent particle density (or if the particles have no closed pores, also the true particle density) can be measured by fluid displacement methods, i.e. pyknometry, which are in common use in industry today. The displacement can be measured with either liquids or gases and there are, therefore, two groups of techniques and instruments available, as follows.

2.4.1.1 Liquid Pyknometry There are several British Standards that deal with liquid pyknometry applied to specific materials I2 •13 ,14,15,16. A pyknometer bottle of up to 50 ml volume is usually sufficient for fine powders but coarse materials may require larger calibrated vessels. BS 1377:1975, for example, requires the use of a 1 litre cylindrical gas jar, closed by a ground plate, to measure the density of soils that contain particles coarser than 2 mm BS test sieve (but not coarser than 37.5 mm). Fig. 2 shows a schematic diagram of the sequence of events involved in measuring particle density using a liquid pyknometer; the density is clearly the net weight of dry powder divided by the net volume of the powder calculated from the volume of the bottle less the volume of the added liquid. If distilled water is not suitable because it dissolves, reacts with or is absorbed by the solid, another liquid must be used. 2.4.1.2 Air Pyknometry This is a method based on displacement of air or other gas. The procedure and principle described here are explained with reference to a particular commercial instrument (Beckman Model 930 Air Comparison Pyknometer) but the concept is general, applicable to other products.

18


• glass stopper capillary Weigh empty bottle

Liquid pyknometer

Fi II about '13 with powder and weigh

Drive off bubbles by evacuation, ultrasound or bOiling

Add liquid up to almost full

Top up and weigh

FIG. 2. Principle of and sequence of events in liquid pyknometry.

An air comparison pyknometer usually consists of two cylinders and two pistons, as shown in Fig. 3. One is a reference cylinder, which is always empty, and the other has a facility for inserting a cup with a sample of the powder. With no sample present, the

19

PROPERTIES DEPENDENT ON SINGLE PARTICLE STOP

STOP

-I

-

I

-.Jo.-

,,

,,

I END

> _____ MEASURING PISTON

START

Start of test

I

II!

~

CUP

i "

"

i

I

star'l. No

-.ToPistons moving (Pressure balance is ~aintained)

fl

• ,,

,,

,, ,

I

I I

I

! End of test

I

'---

.i I

f""T'T I I

lIolum e in ml

'----' sample volume

FIG.

3. Principle of and sequence of events in air pyknometry.

volume in each cylinder is the same so that, if the connecting valve is closed and one of the pistons is moved, the change must be duplicated by an identical stroke in the other in order to maintain the same pressure on each side of the differential pressure indicator shown in Fig. 3. If a sample is introduced in cylinder B, and the piston in cylinder A is advanced all the way to the stop, to equalize the pressures the measuring piston will have to be moved by a smaller distance because of the extra volume occupied by the sample. The difference in the distance covered by the two pistons, which is proportional to the sample volume, can be calibrated to read directly in cubic centimeters, usually with a digital counter. The method will measure the true particle density if the particles have no closed pores or the apparent particle density if there are any closed pores, because the volume measured normally excludes

20


any open pores. If, however, the open pores are filled either by wax impregnation or by adding water, the method will also measure the envelope volume. The difference between the two volumes measured will yield the open pore volume which is sometimes used as a measure of porosity. Measurements of volumes also provide a useful way of establishing shrinkage due to a physical or chemical change. Materials that are not surface-active and are incompressible can be measured using room air, with no additional connections necessary, and following the standard procedure which exposes the sample to air pressures from 1 to 2 bar. Compressible materials may have to be tested at lower pressures, from 1/2 to 1 bar, whilst surface-reactive materials, which adsorb some or all of the air constituents, may be tested using helium (or another inert gas) instead of air. 2.4.2 Measurement of Effective (Aerodynamic) Particle Density The effective particle density is based on the average density within an aerodynamic envelope around the particle; any open or closed pores are included, therefore, in the volume measured. One obvious way to measure the volume of the open pores is with a mercury porosimeter but this is only suitable for coarse solids and the necessary equipment is very expensive. There are four other, simpler methods used for measuring the effective particle density, as follows. 2.4.2.1 Caking End Point Method In the petroleum industry, the effective particle density of freeflowing cracking catalysts is measured indirectly by measuring the open pore volume. This consists of adding water or another liquid of low viscosity and volatility, to the powder until the liquid has filled all the open pores and it starts coating the external surfaces of particles: the powder cakes-up by surface tension at this point and stops flowing. If the volume of water needed to just cake the powder is x cm 3/g (assumed to be the pore volume) andpa is the true particle density


21

as measured by an air pyknometer or a wet pyknometer, the effective particle density is as follows:

Pp = l/[(l/Pa) + x] A somewhat similar method is described in British Standard 812 which is specific to mineral aggregates, sands and fillers l 4, i.e. fairly coarse particles that do not react with water. This requires an open-top, wire mesh basket, together with a tray slightly larger in diameter than the base of the basket. The sample is put in the basket, immersed in water and left for 24 hours; it is then weighed under water (weight B). The bulk solid is emptied into dry cloth and the basket is weighed empty, again in water (weight C). The solid is dried but only until all visible films of water are just removed but it still has a damp appearance, and it is then weighed (weight A). The drying process is then completed in an oven and the solid weighed again (weight D). The weighing under water gives the weight of solid less the weight of displaced water (B - C), and the effective relative density is calculated as D/[A - (B - C)]. This can be shown to be equivalent to the above equation for the caking end point method. The same procedure can be used for a rough estimate of the apparent relative density if weight D is used instead of A (and the corresponding step in the procedure to determine A is omitted).

2.4.2.2 Bed Voidage Method The bed voidage method was developed by Abrahamsen and Geldart ll as a simple alternative to the caking end point method outlined above, when the end point cannot be clearly identified. This method is based on an observation that the minimum packed void age is virtually the same for particles of similar size and particle shape. The procedure is that between 0.2 and 0.25 kg of a control powder c of known particle effective density Ppc is poured into a measuring cylinder and tapped until it reaches its minimum volume, corresponding to the maximum bulk density Pbc- This procedure is repeated with the unknown powder x (ideally, several control powders should be used). If the porosity after tapping is assumed to be the same for both powders c and

22


x, then their effective densities are in the same ratio as their bulk densities. As it is not always possible to find known powders of the same shape as the test powder, Abrahamsen and Geldart ll introduced an experimentally-determined factor k which varied from 0.82 to 1.22 for conversion from spherical to angular or the other way round. They tested the method using non-porous powders and compared it with liquid pyknometer results, and found an excellent agreement.

2.4.2.3 Bed Pressure Drop Method This is a modification of the technique originally proposed by Ergun17 and developed by Geldart 18 . It is particularly useful for powders which are compressible, that is those which have a large difference between their aerated and tapped densities (refer to section on bulk powder density for definitions). It is based on making measurements of bed pressure drop as a function of gas velocity at two voidages, when gas is passed through the bed of powder in the laminar flow regime. The bed is first fluidized to get it well mixed and then gently settled so that it assumes its maximum voidage. Pressure drop is measured for at least four velocities; the bed is then tapped so that it reaches as Iowa voidage as possible and more measurements are made. If the first set of measurements is given subscript 1 and the second subscript 2, and if s stands for the gradient of pressure drop with gas velocity, Pb is bulk density and Pp is the particle effective density, it can be shown (using the well-known CarmanKozeny equation) that the basic relationship is as follows:

SJS2 = (PbJPb2) [(Pp - Pb2)/(Pp - Pbl)P In this equation, Pp is the only unknown and can be readily found, by trial and error. Note: The above equation assumes that the constant in the Carman-Kozeny equation is the same for both maximum and minimum voidages. This is not quite true since the 'tortuosity factor' is slightly different; however, the error induced is small (less than 10%).


23

2.4.2.4 Sand Displacement Method Another possible way of measuring the envelope density of coarse particles is by the sand displacement method. It uses fine sand, into which a known amount of the coarse particles of the sample is mixed. The density of the sample is determined from the difference of the bulk density of the sand alone and that of the mixture. This method is sometimes used for density determinations of coarse bone particles, for example, and it gives lower density than that of the solid bone as measured by pyknometry. A similar method has recently been used by Buczek and Geldart19 • Comment on Effective Particle Density The use of the effective (aerodynamic) particle density is largely restricted to fluidization and pneumatic conveying applications at the moment. It is, however, potentially useful in other areas like particle size measurement by sedimentation or elutriation, or flow through packed beds where it has not yet been fully accepted. 2.5 SURFACE AREA OF POWDERS One of the most important characteristics of fine powders is the area of the surface of the solids; this is usually expressed as specific surface area. This is a measure of the fineness of the powder (the specific surface increases with decreasing particle size) as well as of its porosity but it is unable to discriminate between a monodisperse powder and one containing a wide range of particle sizes. Surface area is important in all applications where the process is surface-dependent like in mass and heat transfer, flow through packed beds or fluidization. Activity of drugs, setting time of cement and effectiveness of cracking catalysts are just three examples of direct dependence on specific surface. Some such materials, like fillers or catalysts, are often specified in units of specific surface rather than in particle size and its distribution. Specific surface also offers some practical advantages, in favourable cases, in the ease and speed of measurement and also in that it gives a

24


single parameter for the tabulation and analysis of the properties of a large number of related experimental or production samples. There are many methods of surface area measurement: any surface-dependent phenomenon can be used for such measurement. The following is an incomplete list of some of the available methods:

* permeametry .* gas adsorption * dynamic gas adsorption * gas diffusion * hindered settling * adsorption from solution

* flow microcalorimetry * dye adsorption * porosimetry The different methods rarely agree because the values obtained depend on the procedures used and also on the assumptions made in the theory relating the specific surface to the phenomena measured. The actual method to be used for surface area measurement is selected depending on the purpose of the measurement. The specific surface determined by any of the above methods can be converted into an equivalent mean spherical diameter Xsv using the following simple equation: Xsv = 6/Sv , where Sv is volume specific surface. The equivalent mean diameter Xsv is the size of a spherical particle which, if the powder consisted of only such particles, would have specific surface area the same as the actual sample. Only the first two methods are discussed briefly here because they are most widely used in industry; for the others, the reader is referred to one of the standard textbooks available on the subjectS3 • Permeametry

The basis of this method is the measurement of permeability of a packed bed of powder to a laminar gas flow. The determination of permeability can be made either at continuous, steady-state

PROPERTIES DEPENDENT ON SINGLE PARTICLE approach velocity

pressure drop

25

porosity (voidage)

• AP

u=----

f viscosity

FIG.

t bed depth

weight specific surface

4. Carman-Kozeny equation.

flow (at constant pressure drop) or at variable flow (constant volume). The Lea and Nurse apparatus and the Fisher Sub-sieve Sizer are examples of the constant flow instruments while Griffin and George permeameter (oil suction) and Reynolds and Branson (mercury suction) are examples of the variable flow instruments. All of the permeametry methods are based on the CarmanKozeny equation given in Fig. 4 which relates the approach velocity u to the porosity of the powder e and the specific surface of the sample Sw' The specific surface calculated involves only the walls of the pores of the bed which are swept by the flow and it does not take into account the pores within the particles which do not contribute to the flow. The surface measured, therefore, is an envelope surface area and it can be very much smaller than the total surface area of the particles as measured, say, by gas adsorption. The Lea and Nurse apparatus, shown schematically in Fig. 5 as an example of a permeability method, was the first equipment designed for routine service. It is now quoted in some national standards2o ,21 including an International Standard now in preparation (ISO/TC 119/SC/2 N 279). The sample is compressed to a known porosity in the permeability cell and maintained within it by means of porous paper discs at each end, and supported by a rigid perforated plate. The standard cell size of cross-sectional area of 1.207 cm2 is usually used. Dry air flows through the bed (drawn by an aspirator or a pump) at constant rate and then passes through a capillary which serves

26


==-===:_

dry air _--- 2 (for specimens formed of 10 increments) but the value of 3 1/2 was used in the rest of the study. 3. For specimens formed by the method proposed, the unconfined yield stress is uniquely determined by its average bulk density, independent of the number of increments in which it is formed. 4. The results obtained for the failure function of the Ti02 used in the study agreed closely with that obtained using a lenike shear cell. 5. For most powders the stress needed to form a satisfactory compact is higher than that which will occur in a storage hopper but it is within the range of stresses that are likely to occur in

58


other equipment such as screw conveyors, extrusion presses and compaction dies.

Comments: This method of testing the failure function in a uniaxial compaction test clearly represents a simplification compared with the procedure required by the lenike shear test. The equipment used in the original study included a constant strain rate device for loading the test specimen. If this test were to be adopted for general use (and it certainly has such potential), the loading of the specimen could be done with a much simpler arrangement like the system of adding weights used in the Compression Tackiness Tester described in the following section. Another simplification would be to use fewer sections, say between 5 and 10, the length-to-diameter ratio LID could be just 2 and the mould could be made of a material which is easier to machine like brass or aluminium alloy. 4.4.2.2 Compression Tackiness Tester This is a simple version of the uniaxial compression test, first developed by Colgate-Palmolive25 but now used by other industries. The procedure described below is based on an article by Monick25 but with slight modifications, as practised at one large chemical company. The method is an analogue of measuring tackiness of a powder by squeezing it by hand and observing if the mass breaks apart readily or else remains as a lump. The equipment (Fig. 22 and 23) consists of a piece of pipe split lengthwise into two halves to form a mould, a clamp, a number of Micarta disks (t in = 6 mm thickness, weighing 20z = 57 g and 1 in = 13 mm, weighing 40z = 113 g) a stand and a rod with aluminium disks attached at either end (thUS forming a compression plunger) for applying pressure to the powder sample. The pipe is 31 in (89 mm) Schedule 40, made of nickel-plated brass. The clamp is made of a strip of flat metal, it fits around the slit mould and is tightened by means of a wing nut. When testing a sample of powder, the mould is placed on the base of the stand, with the two halves clamped together, the

NON-AERATED FLOW AND HANDLING PROPERTIES

,

I'

I

FRONT ELEVATION

I I

,

I'

I I

I I I

I:

I

II

d

I: I'il

I I

"

,I

I

PLANVIEW -

,

d

I I I I

59

I I I I I

:

-

FIG. 22. Mould for compression tests.

..

PLUNG ER

I

rn,

I I

II L .

_ _ _ STAND

I I

,

I

I

SPLIT MOULD_

CLAMP

I:, ,,

H-: ~ : I I

I

FIG.

,: I I

I I I I I

:I I

I

I

I I

I

I

j

\

23. Compression tackiness tester.

mould is filled with powder, and the surface is scraped flat. One

! in Micarta disk is placed over the powder surface and the bottom of the compression plunger is lowered onto the Micarta disk. In addition to the total weight of the plunger and the Micarta disk, further weights are placed on the top disk of the plunger (usually 22 lb = 10 kg in total) and this weight is allowed to remain for 2 minutes. The compression weights are then removed, plunger

60


lifted and locked in an off position, the clamp around the mould is removed and the mould is carefully pulled apart. If the cylindrical briquette of powder does not collapse under the weight of one disk, add another! in (6 mm) Micarta disk, then the k in (13 mm) disk, and if the briquette still stands after 30 seconds, add the weight ofthe plunger too. In 30-second intervals, add weights on top of the plunger in half-pound (227 g) increments until the briquette collapses and note the total load at failure. The total compression force of 221b (10 kgf) is supposed to simulate pressure greater than that at the bottom of a 6 ft (1829 mm) high pile of bags in storage. While the single value at 221b (10 kgf) may be used to describe "tackiness", (higher initial loads should be used with non-tacky products like inorganic powders, or lower values with tacky, organic materials), it is useful to repeat the tests at several loads to see the effect of the compression on the stress at failure ("tackiness value") because the shape of the curve is also an indication of tackiness: non-tacky products give essentially a horizontal line while tacky products show the effect of increasing compression. As would be expected, moisture increases tackiness, so does· the dust content (or fineness of the powder) and, with some powders like, for example, detergents, tackiness increases with temperature. Note that a portable tester based on the same principle as the compression tackiness tester is defined by a British Standard for testing of soils 16 • It uses a split mould 38 mm in diameter and 76mm long (note the 2:1 ratio), with the sample of soil extruded into it from a sampling tube. The load is applied manually by a rotary handle and a lead screw, through a calibrated spring. Comments on the Compression Tackiness Tester: This is clearly a fingerprinting method of a proven practical value. The result does not represent, however, the unconfined yield stress corresponding to the compression load because the compression stress varies within the height of the briquette. This is because the load during the sample compression is partly taken up by wall friction and the stress (and the bulk density) therefore reduces in


61

the do.wnward directio.n. This, ho.wever, do.es no.t matter as lo.ng as the test co.nditio.ns and materials o.f co.nstructio.n are co.nstant, and o.nly relative co.mpariso.ns o.f po.wder characteristics are so.ught. There is, ho.wever, a po.int to' be made abo.ut the cho.ice o.f the dimensio.ns in the tester. As was reported in the previo.us sectio.n, Williams et al. 24 have sho.wn that the length to' diameter ratio. o.f the briquette sho.uld be equal to. or greater than 2, and this co.nditio.n is no.t satisfied in this test. The reaso.n fo.r this requirement is that the briquettes do. no.t fail alo.ng slip planes at 45 degrees but at angles greater than that; the briquette must be o.f sufficient length so. that the slip plane do.es no.t intersect either o.f the end plattens because if it do.es, part o.f the specimen has to' slip alo.ng the end platten during failure, and this requires additio.nal wo.rk. The crushing strength o.f cylindrical specimens is, therefo.re, high fo.r lo.w values o.f LID (length-to.-diameter ratio.) and beco.mes independent o.f LID abo.ve the critical value 24 • The LID ratio. in the Co.mpressio.n Tackiness Tester is o.nly 1.43 and the strength o.f so.me o.f the po.wders under test can be increased by this. Increasing the length o.f the briquette wo.uld, ho.wever, further reduce the co.mpressio.n stress at the bo.tto.m o.f the briquette during its preparatio.n and the LID ratio. used here is an inevitable co.mpro.mise. 4.4.2.3 Large-scale Uniaxial Test There has been a report from Japan recently31, o.f a large-scale versio.n o.f the uniaxial test, applicable to. co.arse granular materials like co.al and limesto.ne co.ntaining particles up to. 50 mm in size. Altho.ugh such co.arse materials are o.utside the sco.pe o.f this repo.rt, the test still deserves a mentio.n here because it is a lo.gical extensio.n o.f the tests repo.rted in the two. previo.us sectio.ns. The paper do.es no.t describe the sample preparatio.n in much detail but it appears that the test mo.uld is a split cell like the o.ne used with the Co.mpressio.n Tackiness Tester except that the LjD ratio. is always equal to. two.. The mo.uld is no.t split into. sectio.n ho.rizo.ntally like the o.ne used by Williams et al. 24 and the material is repo.rted to. be filled in the mo.uld in three layers. Three different

62


sizes of the briquettes used are reported, depending on the maximum particle size of the material to be tested: 300 mm diameter and 600 mm height for maximum particle size of 20 mm and larger, 100 mm diameter and 200 mm height for 100 mm particles, and 50 mm diameter and 100 mm height for 3 mm particles. All test moulds are laminated with Teflon and the compressive force applied is anything up to 392 kPa. The full curve of material strength against the compaction load is tested by reducing the compaction load in steps until the last test piece cannot stand alone. A universal compression testing machine is used for both the sample preparation and the actual testing, with the results recorded automatically by an X-Y recorder. The effects of moisture content and of particle size were studied in the reported investigation 31 but no attempt was made to compare the results with other tests. The strength of the samples not containing the large particles was found greater and this seems to contradict Jenike's statements about only the fines being responsible for failure in powders.

Comments: This again is a test suitable for fingerprinting purposes, or comparisons to show the relative effect of moisture, consolidation time or particle size. The function measured is not the failure function (i.e. unconfined yield strength against the consolidation stress) because the specimens are not homogeneous vertically due to the action of wall friction during sample preparation. Nevertheless, the test is well-worth considering for relative testing of coarse materials. 4.5 TENSILE STRENGTH Tensile strength is a fundamental failure property and it represents the minimum force required to cause separation of a bulk structure without the complications of overlapping boundaries and other disturbances in the failure plane. The tensile strength of a powder is not directly required in


63

equipment design (other than to allow completion of the family of yield loci) but it is a useful measure of the cohesivity of a powder for fingerprinting of products, quality control etc. Recently, Yokayama et al. 32, for example, used the tensile stength of powders in Geldart's groups C and AC in a study offloodability. There are basically two ways of testing tensile strength depending on the direction of pull with respect to the direction of compaction. The split cell testers pull the sample apart at 90 degrees to the direction of compaction whilst the lifting-lid testers pull in the same direction as the direction of the compaction force. The results of the two methods do not agree because the powder compacted in a cell is not isotropic. The results of tensile tests are generally of poor reproducibility and a large number of replicate tests is necessary. Note that there is another way of preparing the sample, by hydrostatic compaction in a triaxial test, but this is a research tool and, as such, outside the scope of this guide. 4.5.1 Split Cell Testers As the name implies, the split cell testers use a cell, usually in the form of a ring similar to that used in the lenike shear cell or the wall friction testers, but both the cell and supporting plate are split vertically, across the diameter. The sample is compacted vertically in the usual way (by application of normal loads via a lid) and the sample is then pulled apart by moving the split halves away. There have been some research devices designed on this principle, for example by Boden33 , who suspended the cell supports on air bearings to minimize friction, but such devices are too expensive for general use. There is a split cell tester available commercially, however, which uses an ingenious mechanical system to counter-balance friction so that the tester is inexpensive and field-portable. The machine is a refinement of the device developed by Warren Spring Laboratories for research use: the assembly has been made more compact and robust, suitable for site, works or lab use. Fig. 24 gives a schematic diagram of the Ajax-WSL Tensile Test Machine: it consists of a stainless steel horizontal circular cell (500 mm diameter, 10 mm deep), split diametrically, one half

64


-

powder loading plug split celt

~

loading ring I

~

clamp

back balance spring

pivot bearing

loading screw

FIG.

24. The Ajax-WSL Tensile Test Machine.

of which is fixed to the main body of the machine and the other, matching half mounted on a pivot block supported by low friction radial bearings. Twin opposing action torque springs are each attached to screw-adjusting blocks so that the moving part of the cell can be positioned in a contact-null balance position against the fixed half. The assembly is mounted in a stainless steel body on which is fitted a clamping screw, to secure the two halves for loading, and a tensioning and adjusting screw assemblies for setting the initial null balance loading on the cell. The moving cell portion may be removed from the machine for cleaning and checking the density of the sample after failure by weighing the contents. The increasing separating force is applied to the cell by means of null balance low hysteresis springs which are tensioned by a lead screw. A counter is directly coupled to the tensioning screw for recording the precise spring extension. A filling cylinder and a loading plug are also supplied with the machine. Four sets of matched springs are supplied to give a range of strength tests. A


65

calibration chart is given to convert the reading of the counter at failure into tensile strength. The following machine preparation and test procedure is recommended by the manufacturer; further details are given in the instructions with the instrument. With the machine on a flat surface, the counter is turned to zero. If in doubt as to the set of springs to be used, use the strongest set and then repeat the test if the results are not sufficiently sensitive. Adjust the back tension screw so that the two halves are just touching, with no excess load holding the cells together. Screw the counter back and then gently turn forward, noting at what count the cells part. Adjust the back screw so that the cells part just as the counter passes through zero. In this setting, the cell is at null balance with no load on the sample until the loading spring is further extended as indicated by the counter reading. The counter is then turned back so that a small force is holding the cell halves together. The cell is then locked firmly in the closed position by the top clamp screw and the loading cylinder is placed squarely home on top of the cell rim. In operation, the sample is first compacted into the cell under a known normal load, using the plunger provided; if the mass of the material compacted into the volume of the cell is known, the density of the sample may be determined-the sample can be weighed after the test by brushing it out of the cell. The top clamp screw is then released and the sample is ready for the strength test. This is accomplished by turning the loading screw to increase the spring length so that the load is imparted to the sample after the counter has passed through the zero position. The load is increased steadily until the sample breaks apart under the applied tensile stress. The reading of the counter is noted at this point and the corresponding stress is read off the calibration curve supplied. Another commercial instrument based on the split cell principle is the Hosokawa Micron Cohetester. One half of the cell is stationary and the other, moveable cell with its mounting frame is suspended from the top. The instrument can also detect expansion of the powder.

66


4.5.2 Lifting-Lid Testers This method is shown schematically in Fig. 25; it is easier to use and, according to Williams35, gives results with less scatter. Williams recommends it for monitoring of cohesivity of a material, or for comparisons. Boden33 describes the system he developed at Bradford and this is briefly described in the following; no commercial instruments based on this principle are known to the author. A mould in the form of a ring, like the lenike shear cell ring for example, is used and a lid which just fits inside it. The base of the cell and the lower face of the lid are covered with sellotape on which a glue is spread. The cell is filled with the powder to be tested and it is scraped level with top of the cell; the lid is placed in position, on top of the sample. A compacting load is applied to the lid and left in position until the glue has hardened. The lid is then slowly lifted via a tensile load cell, by an electric motor. The stress required to break the specimen is noted and both the lid and base of the cell are examined after failure to make sure that both are covered with powder, showing that failure occurred within the powder specimen and not at the surface. If both surfaces are not covered with powder, the test is rejected. It remains to be seen whether this method gains greater popularity than the split cell testers; at the moment, the latter are more widely used simply because there are commercial models available.

lid

base_

""

t

glue covered surfaces

lifting force

powder

FIG.

25. Principle of the lifting-lid tensile tester.


67

4.6 COHESION AND COHESION TESTERS Cohesiveness of a powder is a compound measurement and its fundamental origin lies in the propensity of the individual particles to stick together. There are many different terms used in describing the handling property of sticky or cohesive powders: cohesivity, tackiness, cohesiveness or simply cohesion. Whether the phenomenon is caused by the presence of moisture, electrostatic charge or simply by the greater fineness of the powder which increases the surface area and, with it, the bonding forces which act on the surface, the net effect is that the powder has internal cohesion. This internal, interparticulate cohesion is analogous to the cohesion within solid materials like steel except that, in that case, it is an inter-molecular bonding which is much stronger. The powder cohesion demonstrates itself in the flow and mixing properties of the powder and, more specifically, in the strength that the powder exhibits when subjected to an external tensile stress. It could be argued, therefore, that the only direct way of measuring cohesion is via the tensile strength (see section 4.5). In shear, the cohesiveness of the powder demonstrates itself in the shear strength of a consolidated sample which is not under any external normal load during the test: the internal cohesion force, however, is equivalent to a normal load and leads to a sizeable shear strength in pure shear (the intercept C of the yield locus with the y-axis in the shear stress-normal stress plot in Fig. 14) which is called rather misleadingly "cohesion". It follows, therefore, that the tensile strength tests reported in section 4.5 should also be included in this section, as a direct way of measuring cohesion. All of the other ways are, therefore, indirect in that they measure other properties related to or as a consequence of the internal cohesion forces. The tensile strength described in section 4.5 is the most direct way of assessing cohesiveness because it measures the internal adhesion properties independent of the mechanical interactions of a shear plane. From the practical point of view, however, such interactions do occur in powder flow and handling and it is perhaps

68


relevant to include them in the measurement. The Cohesion Tester described in the following falls in this category. An instrument which attempts to measure cohesion was originally developed at Warren Springs Laboratory and is now available from Ajax Equipment (Bolton) Ltd. It is designed to aid the assessment of flow properties of bulk solids in that it measures the cohesive strength of samples of powders in varied states of compaction, from lightly settled conditions to firm compacts. It attempts to measure directly cohesion as defined in section 4.1.2, i.e. the shear stress at failure, with no normal load acting upon the surface of failure. These conditions represent circumstances in powder storage and handling where the material is not confined by boundary walls or adjacent masses of product so that, in order to fail, it does not have to overcome externally applied forces or constraints but merely the strength developed by prior compaction of the powder. The equipment is shown schematically in Fig. 26; a radiallyfinned annular, open-ended cell is mounted on a vertical spindle carried in low friction bearings and supported in a "floating head" arrangement whereby the weight of the cell is fully supported by a light spring. The cell has 8 vanes and its, outer and inner diameters are 100 mm and 50 mm respectively. Torque is applied to the cell by means of a null-balanced low hysteresis coil spring extended by a screw spindle and pulling through a toothed belt onto a pulley on the cell axis. A sensitive measure of the stress at failure is given by a totalizing counter mounted on the operating screw. The necessary equipment also includes a sample container, fitted with a removable loading sleeve, a loading bridge and plunger, and a set of weights for compacting the samples. The instrument is mounted on a bench stand and is fitted with a screw height adjustment for offering the mechanism to the sample. The full head can be reversed through 180 degrees for the use of the machine in an overhanging position if required. The sample preparation is similar to that used with other test of failure properties: coarse particles or agglomerates are first removed by sieving or similar means. The sample is then compacted at a known normal load or, alternatively, to a given bulk

NON-AERATED FLOW AND HANDLING PROPERTIES 1. loading

LOADING BRIDGE

COMPACTING LID _ ---~

LOADING RING

4I1----CELL

2. Sample preparation

TORQUE PULLEY

-

FLOATING HEAD-----I..

FIG. 26. Cohesion tester.

LOADING SCREW

69

70


density, including any time-consolidation if required. The loading is with the plunger, onto the powder surface which has been trimmed level. Afterwards, any projecting material is lightly scraped away using a sharp-edged instrument. The sample container may then be weighted with the contents to ascertain the bulk density of the sample, taking account of the volume and tare weight marked on the sample container. In the actual test, test springs are fitted, appropriate to the strength of the sample to be tested. If in doubt, the strongest set is fitted and the test repeated with more sensitive springs afterwards if necessary. The counter is set to zero, with the cell freely settling to a neutral position with springs at each side lightly stretched. The position of the pointer on the pulley at the top of the cell spindle is marked with a felt pen. It is also necessary to note the neutral "floating" position of the cell in terms of the free gap between the cell and the fixed collar of the torque spindle: when this position is restored with the cell embedded in the sample there is no load resting on the sample. With the cell head raised and sample container in place, the main spindle is screwed down to offer the cell to the sample. Continued turning will cause the cell to penetrate into the sample and the movement is continued until the top of the outer rim of the cell head is level with the sample surface. (Note that this condition is specified in the manufacturer's instruction manual: some people recommend deeper penetrations of k or even 1 cm below the sample surface.) The height screw is then reversed to raise the head of the machine whilst leaving the cell embedded in the sample. The gap between the cell and the fixed collar will be increasing and the lifting of the machine head is stopped when the gap becomes the same as the original distance in which the cell was freely suspended. In this condition, the cell weight is completely balanced by the internal spring so that no external force is acting on the sample surface. To fail the sample, the operating handle of the loading screw is turned slowly and both the powder surface and the position of the pointer are observed. An out-of-balance torque is thereby applied to the sample and, as soon as the powder surface is seen


71

to slip, the reading of the counter unit is noted. If the pointer moves from the guide line previously drawn without any clear indication of failure then the sample preparation is suspect and the test has to be repeated with a new sample. The measured count from the counter is readily converted into cohesion strength using the calibration curves provided by the manufacturers. The above described method takes the measurement with the head relaxed and floating; for comparison purposes, the instrument can also be used without relaxing the head, i.e. under load. The Cohesion Tester has been used in industry, mainly as a quality control-type test. It can be useful as an aid to assessment of flow properties and of power requirements in mixers, bulk conveyors and feeders. The original developers of the tester, Warren Spring Laboratories35 , have even tried to correlate the cohesion value with unconfined yield stress determined with the lenike shear cell and found a good correlation for some powders (fc = 6 x C). The tester is, quite obviously, only useful with fine, cohesive powders because: 1. non-cohesive powders have no cohesion, and 2. the cell cannot be pushed into materials which are too coarse.

Experience shows that if the instrument is tried with a noncohesive material like sand, for example, it gives a relative large finite reading although the cohesion should be zero. In addition to the direct measurements of tensile strength and cohesion, there are a number of other, indirect ways of assessing cohesiveness. These are discussed in sections 3, 5.2.4 and 4.8 which deal with categories of powders, Hausner ratio and flowability respectively. 4.7 ANGLE OF REPOSE AND OTHER HANDLING ANGLES The angle of repose is defined as the angle of the free surface of a pile of powder to the horizontal plane. Depending on the conditions under which the pile has been poured and how the

72


angle is measured, somewhat different values of the angle can be obtained for the same powder. The test is only relevant to noncohesive or only slightly cohesive powders which form a unique angle when poured or drained. Various types of angle of repose are used to assess powder flow ability in an empirical manner. There are two main definitions of the angle of repose as follows. The POURED ANGLE of repose is the angle measured on a pile poured freely onto a flat surface while the DRAINED ANGLE of repose is the angle measured on the conical surface of powder in a flat-bottomed container if the powder is allowed to discharge through an orifice in the base. The two values are different (the drained angle being greater than the poured angle) because in the first case the powder sliding or rolling down the slope is separating whilst in the second it is converging. 4.7.1 Angle of Repose of a Heap Each of the above two definitions can be measured in various ways and different industries have adopted their own procedures. The problem is that the angle sometimes depends on the diameter of the pile, giving higher slopes at small diameters (resulting in a peaked tip of the pile) than at large diameters where the sliding particles move more-or-Iess in parallel like in a two-dimensional (linear) pile. Some people in fact pour the powder against an edge and measure the angle of the thus created linear pile and not the angle of the conical pile surface. As to the actual procedure and type of angle adopted, it clearly depends on the application in question and the conditions it serves to simulate. Carr describes36 an angle-of-repose plate which incorporates a protractor, indicator wire and a jarring device (the jarring device is used to measure the angle of fall, see below). Probably the only standard procedure, adopted by some industries, is that described by a British Standards37 for alumina. The apparatus consists of a funnel of glass or polythene, with the stem cut-off square and mounted on a tripod stand (or a rigid bridge,


73

/ \ . J ~ __ :"m -----

FIG.

---

27. Measurement of angle of repose.

see Fig. 27). The bottom end of the funnel stem is 5.2 cm above a flat plate of metal or gloss-painted wood. The flat plate has a series of concentric circles engraved upon its upper surface, aligned with the centre of the funnel stem. The diameters of the circles are such as to give different angles of conical surfaces generated from the bottom end of the funnel stem; the angles in steps of 2 degrees are usually selected and suitably engraved on the corresponding circles. The height of the funnel stem is checked by means of a gauge and the alignment is checked visually with the centre spot of the concentric circles on the base. In the measurement, the sample is poured in a steady stream into the funnel, agitating the material within the funnel and the stem by a piece of wire if necessary. The sample forms a cone on

74


the flat plate and the tip of the cone eventually seals the stem of the funnel, thus preventing any more sample passing through. The poured angle of repose is determined from the graduation mark of the circle corresponding to the circumference of the base of the pile. Four different points of the circumference are usually assessed and averaged to determine the angle. The above-described test apparatus requires only a few grammes of the powder and clearly measures the angle of repose at quite small diameters. Some bulk handling equipment manufacturers regard this small-scale test as meaningless, however, and measure angles of repose with substantially larger amounts of material (10 kg, at least) and correspondingly greater diameters of the resulting powder pile. It is clear that the actual method and scale of experiment must depend on the eventual use of the data measured: the small-scale test is suitable for the angle of repose as a general measure of powder handling but, for the design of processing machinery and storage facilities, a much larger amount of powder should be involved in the test. Standardization of such a large scale test is needed. Jarring or vibration of the supporting plate may change the angle of repose, particularly for aerated materials which are likely to flood. Carr36 has defined an "angle of fall" which represents the angle of repose measured after the base plate of the test device is jarred by dropping a 111 g steel ball bearing from a height of 7 in (178 mm) onto the plate five times. The value of the angle of fall (or the angle of difference, i.e. the difference between the poured angle of repose and the angle of fall) is a measure of the material's potential for flooding, according to Carr. The above-defined angle of fall is a rather poor measure of flood ability , however, and is very rarely used in this country; the definition of the measurement conditions is also poor, particularly of the jarring, and the te5',ted values are certain to depend on the mass and the kind of support of the base plate of the instrument. 4.7.2 Drained Angle of Repose There is no standard procedure or equipment for measuring the drained angle of repose. The measurement is only relevant with


75

non-cohesive materials and is likely to be only used in applications where the aim is to estimate the amounts of powder left in dead spaces in hoppers and silos after discharge. The angle is expected to be affected by the degree of consolidation of the powder and its value, as expected, is usually greater than that of the poured angle of repose. It is clear that the measured values will be affected by the dimensions and size of the container used for the measurement, as well as by the way the powder is charged into it, and there is a need for developing a standard equipment and procedure for this test. 4.7.3 Angle of Slide Closely related to the drained angle of repose is the angle of slide38 , as used by some people. This is the minimum angle to the horizontal, of a flat, inclined surface that will allow bulk solid to flow from rest under its own weight. This angle is supposed to be useful in designing stationary chutes but its measurement has not been described as standardized in any way; the value measured is expected to be highly influenced by the material of construction of the chute (as in the case of the angle of wall friction), the amount of material on the chute (bed depth in particular) as well as by humidity (as with most other handling properties). A recent paper by Augenstein and Hogg39 shows how the motion of individual particles depends strongly on the nature of the surface over which they flow: highly roughened surfaces cause shear within the flowing stream while smooth ones cause slip at the surface. Empirical correlations have been obtained which might be useful for design of slides and chutes. 4.7.4 Conveying Angle Very closely related to the angle of slide is the maximum angle of rise (also called "conveying angle") at which a smooth belt conveyor can operate without the bulk powder running back down. There is a need for a standard test method to be developed for this or at least some validation work to be done to relate the angle of rise to the poured angle of repose and angle of wall friction (or any other suitable property) to allow designers of belt

76


conveying systems to have a reliable design criterion. A rough guide, given in the literature 38 , is that the angle of rise used in such systems should not exceed about half the poured angle of repose. 4.7.5 Angle of Sliding Another property described as the angle of sliding38 is defined as the angle to the horizontal, separating substantially stationary powder from flowing powder in flow from a flat-bottom hopper, measured close to the discharge aperture. It is not clear how this angle is actually measured; if it is measured after the flowing material has gone, it then becomes the drained angle of repose. In any case, this angle can only be useful and measured in the actual field conditions and process hoppers, and it is unsuitable as a general powder characteristic because it is affected by the state of compaction and other conditions in the hopper. 4.7.6 Angle of Spatula Closely related to the angle of repose is the angle of spatula which is measured by spooning out powder with a 7/8 in (22 mm) wide flat-blade spatula. According to Carr36 , the spatula is inserted into the bulk powder parallel to the bottom of the container and then lifted straight up and out of the powder, whilst keeping it parallel with a horizontal plane. The spatula may then be vibrated or tapped ina specified manner and the angle of the tangent to the powder surface at the edge of the spatula is measured, to the horizontal plane. The average angle of several measurements is then taken and except for very free-flowing materials, its value is always greater than that of the poured angle of repose (because, once again, it is taken from a consolidated material). The angle of spatula is a simple measure of the flowability of the powder and, according to Carr36 , for a material to be considered free-flowing, its angle of spatula should be under 40 degrees. Materials that have a wide range of particle size and shape may require the use of a broader blade of spatula than the width specified above.


77

The angle of spatula is clearly useful as a very simple and rough measure of flow ability but the errors in measurement are bound to be very large. The angles of repose and of spatula vary from 25 degrees for free-flowing materials right up to 90 degrees or more for cohesive powders; clearly, towards the top end of the range, the measurements become meaningless and widely scattered, making the test inapplicable to very cohesive powders. Conclusions regarding handling angles There is a need for development of the recommended test procedures for the following angles:

* large scale angle of repose * drained angle of repose * angle of slide for the design of stationary chutes * angle of rise for the design of belt conveyors 4.8 FLOWABILITY AND FLOWRATE TESTS This is another rather vague term used, meaning any measure of the ease with which a bulk powder flows out of a hopper or down a chute 38 • Any of the non-aerated flow and handling properties could go under this category but this section only deals with those measures of flowability not covered elsewhere in this guide. There is, of course, a direct way of testing flowability by measuring the time for a standard amount of powder to be discharged from a specified funnel or hopper. This is sometimes referred to as 'free flow time' , as for example specified by an ASTM standard method 40 for testing metal powders; the funnel used is the same as in the International Standard ISO 3923, Part 1. The ASTM method was used in measuring the flow time of sands in a study of particle morphology (see Ref. 10, Chapter 17). For less easily flowing powders, the funnel may be vibrated to assist the flow and the solid thus may even be tested in a consolidated state. This is done in the Durham Vibrating Cone apparatus 41 in which a metal cone is vibrated in a horizontal plane.

78


Another, this time indirect, method of assessing flowability was proposed by Carr36. This was based on measuring four properties of the powder (angles of repose and spatula, compressibility and uniformity coefficient, all as defined by Carr36) and, by awarding points (out of 25) for each, the sum total was then a measure of flow ability and called "Carr's Index". A commercial piece of equipment is available (Hosokawa Powder Characteristics Tester) for measuring the properties necessary for Carr's Index. Flowability is a difficult thing to measure as it depends very much on the design and characteristics of the delivery device as well as on the powder itself. The procedures for the free flow time are far from standard at the moment and improvement might be needed here. It seems, however, that it is far better to measure flow ability by the failure properties which are functions of the powder only and for which theories exist to be used in equipment design and scale-up. The reader is also referred to the use of the Hausner ratio to describe flowability, described in Section 5.2. As most other powder properties, flow ability is greatly affected by humidity.

5 Packing Properties, Bulk Densities

5.1 POROSITY OF A PACKED BED, VOID RATIO Porosity or voidage of a packed bed of powder is defined as the volume of the voids within the bed (i.e. the volume occupied by air) divided by the total (overall) volume of the bed. The void volume includes the pores within the particles if they are porous. The pores in the particles can, however, be excluded from the definition, particularly if the application is in fluidized beds or flow through packed beds because the pores in the particles are in those cases not available to the flow. When porosity or voidage of a packed bed is quoted, therefore, it is important to state clearly the fact whether it has been measured inclusive or exclusive of open or closed pores. The value of the voidage is clearly much affected by the state of compaction of the powder and some powders can be compacted into a whole range of porosities just like in the case of bulk densities. The two properties are in fact related via the particle density in that, for a unit volume of the bulk powder, there must be the following mass balance: (5.1) where Pb is powder bulk density, Pp is particle density and Pa is air density. As the air density is small relative to the powder density, it can be neglected and the porosity e can thus be calculated simply as 79

80


(5.2) The above equation gives the porosity (= voidage) of the powder and whether or not this includes the pores within the particles depends on the definition of particle density used in this evaluation - see section 2.4. There is a considerable confusion between the voidage defined above and the so-called void ratio often used in soil mechanics; it is defined as .d. volume of the voids ratio VR = . volume of the solIds

VOl

and this gives another form of the mass balance, this time for a unit volume of the solids:

(VR + l)Pb = Pp + Pa VR

(5.3)

and if Pa is again neglected, VR can be calculated from:

VR = (Pp - Pb)/Pb

(5.4)

As can be seen by comparison with equation 5.2, this expression has the bulk density in the denominator instead of the particle density as in the case of voidage. The relationship between voidage e and the void ratio VR is from equations 5.2 and 5.4 as follows:

VR/e = PP/Pb

(5.5)

One fundamental difference between VR and e is that porosity e can never be greater than 1 while void ratio VR can be as high as 1.3 in some cases (oats for example). It is preferred to use porosity whenever possible because its definition is more logical and less prone to confusion (as a volume fraction of the whole). It can be argued, however, that in most granular materials the major portion of the strain on handling or compression is due to interparticle movement and that the deformation results from change in the void volume and not in solid volume change. Some people, therefore, prefer to use the void ratio as it relates the volume of the voids to the constant volume of the solids.

PACKING PROPERTIES, BULK DENSITIES

81

5.1.1 Measurement of Porosity Porosity can be measured indirectly via the particle and bulk densities as described by equation 5.2. This is the same method as currently used in many commercial instruments for surface area measurement by permeability in that a known mass of powder is packed into a known volume (i.e. the bulk density is known) and the porosity is evaluated from the knowledge of particle density. It is also possible to measure porosity by permeametry as described in section 2.5. Finally, there is also the possibility of measuring porosity of bulk powders directly by air or liquid displacement of the voids but this is not likely to be very accurate and there are, to the best of my knowledge, no standard procedures or commercial instruments available for such measurements. In conclusion to this section on porosity of bulk powders, it seems that the indirect method of calculating its value from other powder properties, namely the bulk and particle densities, is adequate for most practical applications. The fact whether or not the porosity should include the internal pores of the particles is taken care of by selection of a suitable method for particle density measurement (section 2.4). Care must be taken to avoid a possible confusion between the definitions of porosity and void ratio.

5.2 BULK DENSITY OF A POWDER The bulk density of a powder is its mass divided by the bulk volume it occupies. The volume includes the spaces between particles as well as, of course, the envelope volumes of the particles themselves. The bulk density should not be confused with particle density which is dealt with in section 2.4. There is some confusion as to whether moisture should be included in the mass of the powder when testing the bulk density: in most cases, it is appropriate to include the moisture whilst in some, the moisture has to be excluded. In order to avoid confusion, the basis should be clearly stated in either case. As to the value of bulk density, this very much depends on the

82


state of the powder, particularly on its state of compaction. Since the bulk is a mixture of air (or other gas) and the solid particles, its density can be anywhere between the density of the two phases involved. Thus, in the order of increasing bulk density, there may be aerated bulk density, poured bulk density, tap density and compacted bulk density; all of these themselves depend on the exact treatment to which the sample was subjected in preparation for the test. The standardization here is, therefore, not so much concerned with the actual measurement, as with the sample preparation. The selection of the definition adopted for measurement depends on the application and the test conditions should simulate the actual conditions to which the powder is subjected in practice. Many people in industry measure bulk density in order to get an estimate of how much volume the powder will occupy in storage or in handling. The simplest procedure to be adopted when no great accuracy is necessary is to fill a container of known volume (usually cylindrical), level-off the surface with a sharp instrument and weigh the powder in the container. The container should be ideally at least 1 litre in volume, with a length-to-diameter ratio of about 2: 1. If the as-poured density is sought, no compaction or tapping is applied but some people recommend leaving the sample for about 10 minutes to settle to its stable density. If tap density is sought, by manual tapping, then the test cylinder is tapped by dropping through a height of about 15 mm ten times. For aerated density, some people use a measuring cylinder containing a known quantity of dry powder and aerate the powder by repeated inversion of the cylinder (with top covered, of course). The volume occupied is then read-off as quickly as possible after the cylinder has been righted. This is, incidentally, also a simple way to measure the de-aeration rate, when the rate of bed collapse is estimated rather than just a single volume reading taken. Much the same applies to the simple measurement of the compacted bulk density: the material is tamped, layer by layer, with a rod of some sort, according to a prescribed procedure. If greater accuracy and better representation of field properties are necessary then there is a whole list of more precisely defined


83

_ _ _ SCREEN COVER

1--+-...,

SPACER RING - - - - - - . . .

____

-'===='?~ .....

SCREEN

. - - - - VIBRATING CHUTE

ill...--: :. :.1: :. .:

---,- --

STATIONARY CHUTE

1-: ':f:' -.-.

r01----i

AIR·BORNEFINES - - -___ ,_.' -

FIG.

.! . . ' - : . : . - - -

STANDARD100CCCUP

CUP LOCATION

28. Aerated bulk density determination.

sample preparation procedures, as described in British and International standards. In the following is a short account of such standard procedures. 5.2.1 Aerated Bulk Density There is some confusion as to what aerated density actually means. Strictly speaking, it should mean that the particles are separated from each other by a film of air and that they, on the whole, are not in direct contact with each other. Some authors, however, interpret the term that it means the bulk density after the powder has been aerated. Such interpretation yields, in fact, the most loosely packed bulk density when, for cohesive materials, the strong inter-particle forces prevent the particles from rolling over each other. This gives an open structure unless there is some other, external energy put into the system (vibration or tapping) which gives the particles an opportunity to re-arrange themselves and form a closer packing. Gas fluidization is an obvious means of aerating powders and the reader is referred to section 7.1 where such tests are discussed in detaiL There is yet another way, not involving fluidization, by

84


which a cohesive powder may be deposited in a test cup by an aeration process. This is based on a method described by Geldart42 using features available on the Nauta-Hosokawa powder tester. Fig. 28 shows the basic features of the equipment required. An assembly of a screen cover, screen, a spacer ring and a chute is attached to a mains-operated vibrator of variable amplitude; the vibrators used in small laboratory feeders are suitable for this. The other, stationary shute is aligned with the centre of a preweighed 100 ml cup. The powder is poured through a vibrating sieve and allowed to fall a fixed height (about 25 cm) through the stationary chute into the cylindrical cup. The amplitude of the vibration is set so that the powder will fill the cup in 20 to 30 seconds. The excess powder is skimmed from the top of the cup using the sharp edge of a knife or a ruler, without disturbing or compacting the loosely-settled powder.

5.2.2 Poured Bulk Density This is probably the most widely measured bulk powder property. Different industries have developed their own ways of standardizing the procedure; many use a sawn-off funnel with a trap door or stop, to pour the powder through, into the measuring container. Some decide on a standard volume, usually a measuring cylinder, to be used for the volume measurement and that varies from 50 to 1000 ml depending on the product and its size; one litre is now accepted by many as preferred. It is also better to use a density cylinder with a 2:1 length to diameter ratio than the much slimmer measuring cylinder. Few people standardize the height of fall as well, realising that the height will affect the compaction state of the powder. Finally, it is also the practice in some industries to allow the powder, after pouring it into the container, to settle to its stable density for about 10 minutes before scraping off the top. Probably the best way to show the procedures and the important variables in measuring the poured bulk density is to summarise the International Standard ISO 3923 which, although being specific to metallic powders, can be also used for other materials. The

85


J.1---r-:4;I:~_-

,-----\ I

STAND.----....

FUNNEL

I

\~fr I 1 I

-- -

-

-

DIA2.5

_ , __l _

OR 5mm

r----r---n - - - - , -

DIAaO I

I I

CUP

I

FIG. 29. ISO 3923/1 . 1979 (E) Funnel method.

standard comes in three parts, depending on the properties of the powder to be measured. Metallic Powders - Determination of Apparent Density ISO 3923 Parts 1 to 3 Part 1: Funnel Method (Fig. 29) This standard is applicable to metallic powders that will flow freely through a 2.5 mm or 5 mm orifice. The necessary equipment consists of a 25 ml cup of an internal diameter of 30 mm, a funnel of specified dimensions (60 degree included angle, orifice 2.5 mm or 5 mm, see the standard for the other dimensions) and a stand with a horizontal, vibration-free base. A balance capable of weighing the test sample to an accuracy of 0.05 g is also required. The procedure is to pour the test sample into the funnel while keeping the orifice closed with a dry finger and then allowing the powder to flow into the cup (falling through a fixed height of 25 mm) until the cup is completely filled. The Standard does allow poking with a 1 mm wire to initiate flow from the funnel, through the top of the funnel, if the powder does not start to flow.

86


After levelling the powder with a straight edge (care should be taken not to compress, disturb or vibrate the cup), the mass of the powder is determined and the density is simply the ratio of the mass of the powder to the inside volume of the cup. Three determinations (each with different sub-sample) are recommended for each material to be tested; the total sample volume necessary for this test is at least 100 ml.

Comments: This method could obviously be applied to other than just metallic powders but the necessary pre-requisite is that the powder is free-flowing so that it will flow through the specified 2.5 mm or 5 mm orifice. The density measured is the bulk density of the material as poured under the specified conditions, and this will probably be lower than other bulk densities like the tap density.

t=====E~3"----

STAND _ _ _

FIG.

SIEVE

- - - - BAFFLE BOX

30. ISO 3923/2 . 1981 (E) Scott volumeter method.

Metallic Powders - Determination of Apparent Density ISO 3923/2-1981(£) Part 2: Scott Volumeter Method (Fig. 30) This standard is applicable to metallic powders that will not flow freely through a 5 mm orifice but for which the more precise

87


method, the oscillating funnel method (Part 3 of this same standard) , cannot be used because the powder may change its properties under vibration.

I

I

I

t--~- -7 \

FUNNEL

I

I

'~tt~/ \

OSCILLATING DEVICE_

..

I

--! ! , ----I-

7.5mm ___ '••' - . II

I I I I I I

I,

I I

I

I

I I_CUP I

L

FIG.

i

"=:1

31. ISO 3923/3 Oscillating funnel method.

The necessary equipment consists of a 25 ml cup of an internal diameter of 30 mm, a two-cone funnel of specified dimensions (large funnel feeding into a smaller one via a cylindrical section and a sieve of aperture size 1.18 mm, see the standard for details), baffle box with four glass baffles and a stand with horizontal, vibration free base. A balance capable of weighing the test sample to an accuracy of 0.05 g is also required. The procedure is to pour the test sample into the funnel allowing the powder to cascade over the baffles and flow into the cup (falling through a fixed height of 20 mm) until the cup is completely filled. The Standard does allow brushing with a soft brush to aid the passage of the powder through the sieve. After levelling the powder with a straight edge (care should be taken not to compress, disturb or vibrate the cup), the mass of the powder is determined and the density is simply the ratio of the mass of the powder to the inside volume of the cup. Three determinations (each with different sub-sample) are rec-

88


om mended for each material to be tested; the total sample volume necessary for this test is at least 100 ml. Comments: This method could obviously be applied to other than just metallic powders but the necessary pre-requisite is that the powder is still reasonably free-flowing so that it will pass through the sieve and cascade down the baffle box into the cup. If light brushing is not sufficient to get the powder through the sieve, the Scott volumeter method cannot be used. The density measured is the bulk density of the material as poured under the specified conditions, and this will probably be lower than other bulk densities like the tap density. The method is somewhat similar to the aerated density measurement described in section 5.2.1 in that it uses a sieve, although the latter does not involve passage through baffles. Metallic Powders - Determination of Apparent Density ISO 3923 Part 3: Oscillating Funnel Method (Fig. 31) This standard has just been reviewed by the Sub Committee ISMj 65/2 and will be published shortly. The summary given below is based on the document TC119 N 542 (85/35035) of BSI. This standard is applicable to powders that will not flow freely through a 5 mm orifice but it should not be applied to powders whose particles may disintegrate during vibration such as agglomerated, fibrous or acicular powders. The necessary equipment consists of a 25 ml cup of an internal diameter of 30 mm, a funnel of specified dimensions (60 degree included angle, orifice 7.5 mm, see the standard for the other dimensions) and a vibrating device operating on a supply frequency of 50 to 60 Hz, capable of oscillating the funnel horizontally at a frequency of 100 to 120 Hz and with an amplitude range (peak-to-peak) of 50 to 100 microns. The procedure is to pour the test sample into the funnel while keeping the orifice closed with a dry finger and then allowing the powder to flow into the cup (falling through a fixed height of 25 mm, with the oscillation switched on) until the cup is completely filled. After levelling the powder with a straight edge (care should


89

be taken not to compress, disturb or vibrate the cup), the mass of the powder is determined and the density is simply the ratio of the mass of the powder to the inside volume of the cup. Three determinations (each with different sub-sample) are recommended for each material to be tested. Comments: This method could obviously be applied to other than just metallic powders but the necessary pre-requisite is that the powder is still reasonably free-flowing so that it will flow through the specified 7.5 mm orifice when vibrated. The density measured is the bulk density of the material as poured under the specified conditions, and this will probably be lower than other bulk densities like the tapped density. 5.2.3 Tap Density Tap density is the bulk density of a powder which has been compacted by tapping or vibration following a specified procedure. When using a manual procedure, this consists of pouring a specified mass or volume of powder into a container (a measuring cylinder is again often used here) and then tapping the container against a hard surface, from a standard height (typically 15 mm), by a specified number oftimes (typically 10). The tapping can be done during loading: ASTM has in fact defined tap density46 as "The apparent density of a powder obtained when the receptacle is tapped or vibrated during loading under specified conditions" . There is a bewildering variety of different procedures in the literature for the manual measurement of tap density, proposed by different people for different kinds of powders. Some recommend a fixed mass of the powder to be used (10 g or 100 g for metal powders for example) while others specify a fixed volume (say 50 cc) or use a special density cup with an extension for filling. There is also a wide variety in the determination of the end point: the number of tappings and also their frequency may be specified, or, probably more appropriately, the sample is tapped until a constant volume has been reached. Although many people in industry measure the tap density by

90


tapping the sample manually, it is best to use a mechanical tapping device so that the conditions of sample preparation are more reproducible. One such instrument is the Hosokawa powder characteristic tester which has a standard cup (100 cc) and a camoperated tapping device which moves the cup upwards and drops it periodically (once in every 1.2 seconds). A cup extension piece has to be fitted and powder added during the sample preparation so that at no time the powder packs below the rim of the cup. After the tapping (180 times = 216 s), excess powder is scraped from the rim of the cup and the bulk density determined by weighing the cup. Kostelnik and Beddow47 suggested and tested a Ro-Tap method for determining tap densities. They argued that, when measuring the tap density to be used in the Hausner ratio (see next section for its definition), its measurement should be consistent with the measurement of aerated bulk density in that it should use the same test cup. They considered the measuring cylinder as inappropriate because the visual determination of volume not only makes it difficult to determine the ~ld point but makes it impossible to determine the volume accurately enough.

I I I I I

, I

I

I I I I

I I I I

I

I!...--~-- ~

:

I

:

_

25cc extension

il :I I

I

I I

_

25cccup

I----t---~I

I

FIG.

32. Tap density cup and extension according to Kostelnik and Beddow.

Their cup, called the Hall test cup and accepted as standard by ASTM, is 25 cc in volume and has another 25 cc extension as shown in Fig. 32. This extension is used to make sure that there


91

is enough powder there to start with so that at least 25 cc of the compacted sample is obtained after tapping (as the largest possible Hausner ratio, i.e. the ratio of the tapped density to aerated bulk density approaches 2). For the measurement, the cup with the extension fitted was filled with the test powder, the excess powder scraped off and the sample was then tapped for a given number of cycles. The tapping was done mechanically using the Tyler RoTap device, fitted with a special Ro-Tap fixture which allowed running multiple tests in parallel. At the end of a test, the cup extension was carefully removed and the excess powder in the lower 25 cc test cup was scraped off in a standard manner. The powder remaining in the lower cup was then weighed and this value divided by the 25 cc volume gave the tap density. Satisfactory results were obtained within five minutes of tapping. The authors claim that such short times of rapid densification of the sample were possible because of the unique combination of the tap and the transverse motion of the Tyler Ro-Tap device; they also tested a vibrational device (PM 25 B) as an alternative for mechanical tapping but found it not satisfactory in that the tap density values it gave were lower than those obtained with the Ro-Tap machine. An ASTM standard48 for catalysts from 0.8 to 4.8 mm in particle size recommends a 250 ml measuring cylinder mounted in a cylinder holder weighing 454 g and tapped with a tapping device similar to the Tap-Pak Volumeter Model No. JEL ST2 by J. Engelsmann A.F. of Ludwigshafen, West Germany. The device consists of a base plate with a worm drive of reduction ratio 15: 1, camshaft speed 250 rpm and tapping stoke travel 3.2 mm. The sample is first dried and between 240 and 250 mm is poured into the cylinder using a funnel. The volume of the powder after 1000 tappings and the original weight give the tap density. An alternative ASTM standard, also for catalysts49 , uses the same measuring cylinder but vibrates a stand and funnel during the filling, which proceeds at a rate between 2 and 3 m1/s.

Conclusion There is yet no standard procedure for determining the tap density

92


of fine powders. When more accurate determinations are sought, the use of a machined cup of a calibrated volume and a mechanical tapping device is recommended. In Britain, the mechanical arrangement similar to the one used by the Hosokawa tester is becoming accepted as a preferred alternative to manual tapping, while in the USA the Tyler Ro-Tap device is widely used. The 25 cc cup used by Kostelnik and Beddow seems a little too small though and, from the point of view of minimizing the wall effects, the 100 cc cup is much preferred. There is a need for a standard procedure and equipment to be defined. 5.2.4 Hausner Ratio The Hausner ratio is not a separate test, but it is simply evaluated from the bulk density tests. It is defined as the ratio of the tap density to the aerated bulk density and it is a very useful characteristic of a powder. The use of this ratio as a characteristic of the friction condition in metal powders was first proposed by Hausner50 ; since then it has been used in the flow characteristics, sieving and compaction of metal powders 51 ,52,53, and also in many aspects of general powder technology. The fluidization characteristics of powders, for example 18 , can be correlated with the Hausner ratio for powders in groups C, AC, and A (see section 3.2 for the categories). It has also been related to the morphological properties of sands 10 and Geldart, Hamby and Wong42 found it virtually independent of relative humidity although both of the bulk densities varied with humidity. The usefulness of the Hausner ratio in characterizing powders re-iterates the conclusion in the section on compaction tests (section 5.3) in that the way particles of a powder pack together affects their behaviour in many processes like fluidization, mixing and handling. Kostelnik and Beddow47 found the Hausner ratio extremely sensitive to particle shape and proposed that it could even be used as an index of powder shape. As it is a composite property, it is also responsive to particle size and it increases with decreasing particle size. There is no doubt that the Hausner ratio, as a factor so simple to measure, has a great potential as a general


93

fingerprinting criterion of powder handling behaviour and more research is needed to relate it to other powder properties. 5.2.5 Compressibility (from Bulk Densities) There is another way of expressing the difference between the tap density and the aerated density, as compressibility defined as follows: Compressibility = 100 (PI - Pb)/Pt (%) Compressibilities in excess of 20% are said to correspond to a tendency for creating bridges in hoppers and powders having values over 40% are very hard to discharge at all. There is no overwhelming case for using compressibility in favour of Hausner ratio, although it too is a dimensionless group of similar significance. The Hausner ratio is much more widely used at the moment. 5.2.6 Compacted Bulk Density Sometimes the density of powder compacted over and above what can be achieved by tapping may be needed. There are no standard procedures for this other than the rather specialized soil compaction tests and density determinations (2.5 and 4.5 kg rammer methods) quoted in a British Standard for soils 16 • The compaction test described in the following section may also yield approximate values of compacted bulk density as it is equipped for measuring the sample height.

5.3 COMPACTION TESTS The compaction characteristics of powders and granular materials also provide interesting insight into bulk behaviour of solids. Bulk materials in general are not linearly elastic and cannot, therefore, be characterized simply by Young's modulus and Poisson's ratio as solid materia~s can. If a powder is compressed, the deformation is primarily plastic and the elastic component is small in comparison.

94


The simplest way to test the compressibility of a bed of solids (whether it is dry or wet) is by the one-dimensional compression test. A cylindrical plug of the powder is compressed axially in a cylinder which confines the sample and prevents lateral strains Normal force

POROUS PLATE

_____ ring

powder

---I~

_ _ _-L_ _ _ _..L_ _ _-'~====..':P~O:;ROUS PLATE FIG. 33. Schematic diagram of a consolidometer.

see Fig. 33 for a schematic diagram. The apparatus is known as a consolidometer or oedometer and it consists of a confining ring, porous plugs on top and bottom of the specimen to allow the medium in the voids (gas or liquid) to escape, a means of loading the sample with an axial force and gauge for measuring the vertical deformation. The stress put on the specimen is not purely onedimensional because of the wall friction but the effect can be minimised by using a floating confining ring fitted with a low friction liner. During the test, the axial stress is increased in small increments and the axial deformation, which may be time-dependent, is recorded. The results may be plotted as an axial stress-strain relationship which is non-linear (the rate of strain increase declines with increasing stress), or as bulk density (or voidage, or void ratio) as a function of the compaction stress. The method described above forms the basis of the standard determination of the one-dimensional consolidation properties of soils for civil engineering purposes16 • The diameter of the cell is not specified other than it should be at least 6 mm smaller than the soil samples available for testing, and the depth of the ring is


95

such that the thickness of the specimen is no more than one-third and not less than one-quarter of the inside diameter of the ring. The porous plates are made of sintered fused aluminium oxide, sintered bronze or similar material and filter papers are used between each plate and the specimen. The Standard lays down other important specifications as to the range and accuracy of the depth gauge and the method of loading the cell in small increments so that the pressures generated range from 10 to 3200 kN/m2• The method described in the Standard is applied to wet soil samples and periods of up to 24 hours are recommended for each of a minimum of 4 loading stages to allow time-consolidation to take place before the reading of the deformation is taken. The Standard recommends plotting of the deformation of the specimen (as thickness of the specimen, as strain or as void ratio) as ordinate on a linear scale and the corresponding applied pressure in kN/m2 as abscissa on a logarithmic scale. From this, the coefficient of volume compressibility or the coefficient of consolidation may be evaluated as specified in the Standard. It has been found in practice, from observations in sample consolidations for other tests, that bulk materials with little or no cohesion stabilise quickly and are not very compressible (the strain stops increasing after a certain compaction stress, until particle breakage occurs) while others continue compacting with the increasing stress. The compaction characteristics of powders could, therefore, be used in powder characterization and classification into groups. There is no standard test procedure specifically designed for powders at present and there is scope for developing one. Validation tests could be carried out to determine whether the behaviour of powders in compression could be used as a general criterion of powder flow characteristics and to determine its relation to other handling properties like cohesivity.

6 Grinding and Strength Properties

This section is concerned with those properties which determine particle breakage and attrition. The powder characteristics which playa role in this can be divided into three groups: 1. Characteristics relating to the resistance to destruction such as the particle strength, the probability of breakage, the massrelated work input or the specific reaction force. 2. Characteristics related to the results of the comminution, i.e. the fragment size distribution, the new specific surface area, degree of agglomeration, etc. 3. Other characteristics as combinations of the previous two groups, such as the grind ability , new surface created per unit work input (energy utilization) and others. The above listed characteristics depend on many primary particle properties and operation conditions, such as the nature of the material (brittle or soft), particle size, kind of loading in the mill, intensity and velocity of loading, the shape and hardness of the loading surfaces and the environmental conditions, namely temperature. In general, comminution is an area which has been widely researched and the various tests used in characterizing powders for qualitative and even quantitative design purposes are fairly well developed in comparison with much of particle technology. In addition to this, particle attrition has been a subject of review commissioned by the BMHB immediately prior to this guide 54 and 97

98


the author of that review has dealt with all the relevant tests not only in attrition but in the whole of comminution. There is little point in duplicating the review here other than just to give a list of the tests with a few main comments and conclusions, and the reader is referred to the above-mentioned attrition review for more details.

6.1 SINGLE PARTICLE STRENGTH Fundamentals of comminution concern themselves with the energetics of particle break-up, propagation of cracks and with the size distribution of the daughter particles generated. Tests on single particles have, therefore, played an important role in comminution research. In such tests, a well defined load is placed on a single particle until it is crushed. This load may be applied by direct compression or by impact. The particle must be fairly large to start with and the tests are more easily applied to tablets, pellets, agglomerates or simply large solid particles. There is a whole range of tests based on compression between parallel plates54 and two standards exist to define the rate of application of the crushing force and the geometry of the test apparatus. There are no standards for single particle impact testing but many different procedures have been used in research; they can be classified into those which shoot the particle as a projectile against an impact surface or those where a stationary particle is hit by a surface moving at high velocity (the latter alternative giving a choice of particle orientation on impact).

6.2 HARDNESS Hardness of the particles is one of the fundamental powder properties which affect comminution characteristics, abrasion or attrition. It is defined38 as the degree of resistance of the surface of a particle to penetration by another body, in direct analogy

GRINDING AND STRENGTH PROPERTIES

99

with the definition of hardness of solid materials. Hardness of metals, for example, is commonly measured with static indenters (Vickers, Brinell by the width of indentation, or Rockwell by the depth) or by dynamic rebound of a standard indenter (Scleroscope hardness test) but these are of limited use in testing individual particles unless the particles are large. Apart from this physical limitation, there is also a danger that, as particle size reduces, its strength also reduces and an attempt to penetrate the particle size surface with, say a diamond tip, may result in its breakage. Hardness is often considered to be relative rather than an absolute property. A qualitative, indirect hardness test exists for powders based on the ability of particles of one material to scratch particles of another. This test is based on the work of Mohs: the Mohs' hardness scale (below) lists ten selected minerals in the order of increasing hardness from talc to diamond so that material of a given Mohs' number cannot scratch any substance of a higher number but will scratch those of lower numbers. Reference books list the Mohs' hardness for many known minerals and a simple scratch test with a "hardness pencil" can be used with any unknown materials (providing there is a large enough surface available for scratching). It is conceivable to reverse the situation in that the scratching is done with a particle (or particles) of the material under test, on a surface made of the reference material but no standard procedure is known to the author. 6.3 GRINDABILITY TESTS There are a few tests of grind ability which directly test the relative ease with which a material may be ground. Such tests use a specific type of machine and operation conditions, and start with a specified size distribution or size range of the raw material. Although the various indices derived from such tests may be of general use, the most frequent applications are in equipment scale-up to a particular type of grinding plant. The rule is that the method selected for grindability testing is one that is most closely

100


Table 1 MOHS' SCALE OF HARDNESS

Hardness Material No. 1 2 3 4 5 6 7 8 9

10

Notes

Can mark paper powdered by finger Talc, graphite Gypsum rock salt Can scratch lead Can scratch finger nail Calcite Fluorspar Can scratch copper coin Apatite Feldspar Can scratch window glass Quartz Can scratch a knife blade Topaz Sapphire, corundum Diamond

related to the type of equipment to be used in the large scale plant. Thus, the Hardgrove test is used for scale-up of ring-ball or ring-roll pulverizers in the coal industry because it uses a miniature ring-ball mill; the Bond ball mill test is more relevant to the scale-up of ball and rod mills used in the mineral industry. Probably the best known grind ability test is the Hardgrove method which is used to measure the Hardgrove index of grindability. This is a relative test developed specifically for testing of coaP6 (although it can be used for other minerals) and carried out with a Hardgrove ring-ball machine. This consists of a stationary grinding bowl with a horizontal track. Eight steel balls (1 in in diameter = 25.4 mm) run in the track, driven by an upper grinding ring which is loaded with a compression force of 64lbf (285 N). Each Hardgrove machine is calibrated with standard reference samples of coal of grindability indices 40, 60, 80 and 110 which are obtainable from the US Bureau of Mines in Pittsburgh or other sources. The index is determined from the amount of fines formed if a standard sample weight (50 g) is subjected to standard grinding conditions. The harder is the coal, the lower is the index number, with 100 corresponding to an original standard soft coal. In sample preparation, one kilogram of minus 4.75 mm coal is


101

stage-ground to give maximum material between 1.18 mm (16 mesh ASTM, 16 mesh BS) and 0.6 mm (30 mesh ASTM, 25 mesh BS), with the fines discarded. Fifty grams of such sized coal is placed in the Hardgrove ring-ball machine and turned for 60 revolutions under standard load. The minus 75 micron (200 mesh) fines formed are measured as W (g) and the grind ability index is determined from a calibration chart or from the following equation: Hardgrove Index = 13 + 6.93 W An interesting modification of the Hardgrove test mill is the Warren Spring Laboratory grind ability testS7 which mounts the Hardgrove grinding chamber on a low friction bearing which registers the torque. This, together with the knowledge of the speed of rotation of the mill, allows the energy consumption to be determined by integration of the power/time curve. If the specific surface area of the feed and of the product is determined, the WSL grindability test makes it possible to evaluate the Rittinger number as the amount of new surface produced per unit of energy consumed. It was shown that 57 the Rittinger number becomes reasonably constant for a number of free-flowing, noncohesive materials after about 50 revolutions of the mill whilst for cohesive materials the minimum number of revolutions may be much greater. Another way of testing grindability of materials is to relate it to the energy required to grind a sample to a specific fineness. The Work Index WI defined by Bond 58 is one such parameter. It is measured57 in a laboratory ball mill 12 x 12 in in size (305 x 305 mm) with a smooth lining; a sample of 1.2 kg of less than 2800 microns is ground in stages at 70 rpm. The undersize material (below a selected test size, Pt) is removed in each stage until the rate of product fines is constant. The grindability is defined as the mean mass of fines in grams produced per revolution, Ge. The Bond Work Index WI is deduced from this quantity using the following equation:

102


where Pso is the size at which 80% of the product is finer and Fso is the size at which 80% of the feed is finer (both in microns). The Work Index is claimed to be capable of predicting the power for a ball mill of 8 ft (2.43 m) in diameter in closed circuit wet grinding. In dry grinding applications, the value of WI has to be multiplied by an empirical constant of 1.3. The Work Index is also supposed to represent the energy in kWh required to reduce one short ton (907 kg) of mineral from a theoretical infinite size to a size distribution at which 80% is smaller than 100 microns. Similarly to the ball mill test, there is also a rod mill grindability test57 which uses a small mill 12 in in diameter (305 mm) by 24 in (610 mm) with wave-type lining and a charge of 33 kg of rods. A particular sequence of level operation with 5 degree tilting in both directions is used during the test (using an automatic drive) and the grind ability is once again determined as the rate of production of fines per revolution. Another ball mill test worth mentioning here is the USBM (United States Bureau of Mines) Ball Mill Test57 which is in some respects similar to the Bond ball mill test. The ball mill is 8 x 8 in (203 x 203 mm = diameter x length) and fitted with three lifter bars mounted longitudinally to the inside wall equally spaced around the cylinder. The charge is 100 balls one inch (25.4 mm) in diameter each and the coal charge is 500 g. The mill is rotated at 40 revs/min and a trip counter is used to count the number of revolutions completed. The operating procedure is designed to simulate the conditions in an air-swept mill in that the fines are removed from the grinding zone as they are produced. The feed sample between 75 and 1400 microns is ground in stages and the number of revolutions of the mill is limited in each stage so that about 10% of the total sample will pass a 75 micron (200 mesh) sieve. This fraction is removed by sieving at the end of each grinding stage and the rest of the load returned to the mill. The grindability is expressed as a Ball Mill Index defined as 50 000 divided by the total number of mill


103

revolutions required to grind the sample to a standard fineness of 80% passing through a 75 microns sieve (200 mesh). There are many other indices and specialized grindability tests such as the Holmes grind ability tests, indices for autogenous comminution or for superfine grinding but these are really beyond the scope of this guide and the reader may want to refer to a previous review of grinding equipment57 • Grindability (in its many different definitions) is a composite material property, and depends on many primary material properties (e.g. particle hardness, bulk and shear moduli of elasticity) as well as its flow properties and other conditions like moisture content, humidity of the atmosphere or material composition (rank or ash content of coal, for example). It also depends on the type of mill used for its evaluation. There have been some attempts made by several authors to find correlations relating different measures of grindability; the reader is referred to the literature for details of these 54 • Finally, it should be just mentioned here that, in studying the grinding process and designing grinding plant, it is not only necessary to know the rate at which particles break and the power required to achieve such breakage but also to predict the full size distribution of the product. This can be done using the concept of the selection and breakage functions which define the specific rate of breakage of different sizes in the feed and the size distribution of the daughter particles generated in breaking the feed particles. The measurement of these two functions is much more complex than the simple grindability tests described above and is beyond the scope of this guide. 6.4 IMPACT TESTS Breakdown of particles on impact can be tested either on single particles or on a quantity of the bulk solid, and the result is a measure of particle friability. The available tests have been well reviewed in the recent attrition report by the Board54 and only a mention of the main methods is given here.

104


Leaving aside single particle impact tests which fall into section 6.1, the most common type of multiparticle impact tests are drop shatter tests in which a specified quantity of the test material is dropped through a specified height onto a hard surface or into a container. National and international standards exist on drop shatter tests for coke and coal and similar tests are also applied to iron ores and sinters. Drop shatter tests are in general used on coarser solids than those within the scope of this guide: the fine fraction is in fact often removed from the bulk material before the test. With harder materials like aggregates, sands and fillers, an impact test machine is used59 which uses a 14 kg hammer to drop from a height of 380 mm onto a specified quantity of the sample in a cup. The amount of fines produced by the impact IS an indication of the shatter resistance.

6.5 COMPRESSION TESTS Breakdown by confined compression can be tested in a compression cell similar to that in compaction tests, except that the normal forces used are greater here. A British Standard method59 packs a sample of specified quantity and size grade into a steel cylinder, 150 mm nominal diameter; a plunger is inserted into the open end of the cylinder and the whole compressed in a compression testing machine. The load is slowly increased up to 400 kN and the product is then sized. A similar test has been used in assessing friability of pellets and coal.

6.6 VIBRATION TESTS Friability of tablets or granules is tested quite commonly by vibration in a container or on a sieve. Much the same procedure has also been used in testing finer materials like catalysts, bone char or fertilizers 54 but no standard exists. Some fertilizer manu-


105

facturers in their product quality tests include steel balls with the sample on the testing screen.

6.7 ABRASION TESTS Abrasiveness of bulk solids, i.e. their ability to abrade or wear surfaces with which they come into contact, can be assessed in several different ways. It can be implied from the relative hardness of the particles and the surface with which they are in contact, using Mohs' hardness scale. It can also be described by an Abrasion Index60 which, as a characteristic number, combines the effects of particle hardness, shape, size distribution and bulk density in one factor, independent of the natUI:e of the contacting surface. Abrasiveness, however, is another composite property that is difficult to predict from other properties of the powder and of the material of construction, and it has to be measured directly using standard test procedures. The best way of assessing abrasiveness is to use the actual bulk material and the contact surfaces in question. One such test was developed by Yancey, Greer and Price61 and has been adopted in a British Standard for coke and coal62 • From the test, another "abrasiveness index" is determined by measuring the wear on a standard surface when the surface is brought into moving and intimate contact with the coal under specified conditions. In each test, a sample of 4 kg coal (minus 4.75 mm in size) is processed in specially designed apparatus developed by Yancey and co-workers61 • It consists of a pot 8 in (203 mm) in diameter and 9 in (229 mm) deep with a four-arm stirrer fitted with wearing plates. The stirrer rotates with a small clearance of 5 mm from the sides and bottom of the pot. In the test, the coal is placed in the pot and the stirrer rotated for 12000 revolutions at a speed of 1500 rpm. The wearing plates are weighed before and after the test, and the loss of mass in milligrams determines the abrasiveness index.

106


The above test, although developed specifically for coke and coal, can be used with other powders. 6.8 FRIABILITY/ATTRITION TESTS Friability and attrition have been the subject of a recent review commissioned by the Board54 so that only a summary of the main conclusions is appropriate here. Friability is defined38 as the tendency for particles to breakdown in size during storage and handling, while attrition is the actual, unwanted breakdown of particles. Whilst the above two definitions imply total breakdown, attrition usually means particles getting smaller due to their corners or surface irregularities being knocked off. Attrition is a serious yet little understood problem in bulk solids handling and there have been several initiatives and research projects started recently to further its understanding. Attrition of solids is known to take place during handling or on collection in gas cyclones, for example, but little is known of how it is related to particle properties, which particle sizes are most affected, and how attrition can be related to abrasion. Clearly, large particles are more likely to be affected by attrition; finer fractions are generated by knocking off corners or by complete breakage of the larger particles. Attrition is most detectable in re-circulating systems such as fluidized beds where cyclones are used to return the carry-over material back to the bed. The complete inventory of the bed may pass through the cyclones many times per hour and the effect of attrition is thus magnified many times and therefore easily measured. In such systems, the fresh make-up material added is usually coarser than the equilibrium bed material in order to maintain the optimum, finer size distribution of the solids in the bed. Even in single pass systems, however, attrition is sometimes detected and it demonstrates itself by severe discrepancies in a mass balance of different particle size fractions around the cyclone. It is not uncommon to find that what is supposed to be


107

the coarse product, i.e. the collected solids, is in fact just or nearly as fine as the fine product in the gas overflow. In isolated cases, when the feed solids are very fragile or have a shape which is easily broken (thin-shelled hollow spheres, scale particles or such like), the coarse product may even be finer than the solids remaining in the gas flow and reporting to overflow. Any meaningful evaluations of grade efficiency curves or any other assessments of the separation performance of the cyclone are then impossible. There is clearly a need to investigate the mechanism of attrition to relate it to the fracture properties of the solids, and to develop a realistic attrition "index", similar to that used for abrasion in cyclones. Such an index would indicate the relative importance of operating conditions and design variables such as inlet velocity, feed solids concentration or cyclone diameter. This could then be used in scale-up to predict (or minimize) the effect of the shape, the particle size distribution or the hardness and strength of the feed solids, if known, may allow such predictions without any experimental tests. Generally, better understanding of attrition and its relation to abrasion may lead to better equipment design and operation. Apart from the impact and vibration tests dealt with previously, there are several other tests of friability, some of which appear in British Standards dealing with specific materials. Such tests include mixer tests, jet impingement tests, shear tests, tumbler tests and fluidized bed tests. Of these, the shear tests and tumbler tests are only considered worth mentioning here, with more details available from another BMHB pUblication54 • The shear cells as used in testing yield strength of solids may also be used for testing friability. As large strain is required in order to produce significant attrition, the annular shear cells (which permit infinite strain) are usually used. A variant on the annular shear cell for attrition testing is available commercially from Ajax Equipment in Manchester. Tumbler tests, and more specifically drum tests have been probably most popular and there are at least two British standards describing the procedures, one for testing of coal and coke 15 (the "MICUM" test) and another for iron ore63 , with some American standards also available (ASTM D3402, D4058-81 and E278).

108


The MICUM test uses a drum, into which a bulk solid is placed, of specified quantity and size distribution. The drum is rotated for a given number of revolutions and the product is then analysed for particle size. Internal flights in the drum make sure that the bulk solid is carried round and tumbles around (like in a concrete mixer). Various tumbler and abrasion indices have been proposed to describe friability from tumbler tests54 • The so-called halfMicum tests according to BS 1016, Part 13 (similar to ISO 556)15 uses a 1000 mm diameter drum (internal diameter), 500 mm long, with 4 lifting flights and a door. 25 kg of coal down to 20 mm in size is charged into the drum and the drum is turned for 100 revolutions at 25 rpm. The dust is allowed to settle for 1 minute and the charge is then taken out for particle size analysis. The Micum indices are the fractions in percentages retained on a 40 mm test sieve and passing a 10 mm sieve after 100 revolutions. The Irsid indices are percentages on a 40 mm sieve after 500 revolutions. Extended Micum tests are also defined, when the number of revolutions is greater than 100, up to 900. An ASTM standard for catalysts64 , however, uses a smaller drum (254 mm in diameter, 152 mm long) with a single radial baffle and a charge of 100 g of dried catalyst coarser than 850 microns. The test is for 1800 revolutions at 60 rpm and the loss on attrition is the amount of fines passing 850 microns divided by the original charge weight (100 g). Another tumbler is standardized for testing the resistance to degradation by impact and abrasion for pellets, sinters and sized iron ores63 • The drum used is the same size as the one used in the half-Micum tests except that it has just two lifters longitudinally, at 1800 around the drum. The charge is 15 kg of ore between 10 and 40 mm (prepared according to ISO 3083) and the test runs for 200 revolutions at 25 rpm. The tumbler index is defined as mass of the product retained on a 6.3 mm sieve over the total charge, whilst the abrasion index is the mass passing 500 microns over the total charge. It is clear from this short review of some standard tumbler tests that this area is well served by national standards. Each caters for a specific material and has been developed for a specific index


109

used in the particular industry. It would be very difficult to find a universal test procedure and equipment suited for a wide range of products. The national standards together with a considerable range of other tumbler tests reported in the literature54 give a wide choice of possible tests for many materials, usually mineral, and coarser than say 850 microns.

7 Aerated Flow and Handling Properties

The properties of concern in this section are to do with behaviour of powders in an aerated state; this is relevant in gas fluidization, powder transport and handling. Probably the most important tests in this category are those derived from fluidization and the results of such tests are not necessarily restricted to the area of gas fluidization. It should be emphasized that the following notes apply largely to fine powders (Le. groups A and AC in Geldart's classification) .

7.1 FLUIDIZATION TESTS Excluding the aerated bulk density which is dealt within another section, the aerated properties to be tested in a gas fluidized bed are as follows: (after Geldart ll ,28) Minimum velocity

UMF

Minimum bubbling velocity

UMB

Bed expansion Deaeration rates over a range of gas velocities Many fluidized beds are operated at gas velocities considerably above U MB ; however UMF and the ratio UMB/UMF can give important 111

112


information concerning the behaviour of powders less than 100 . .. mIcrons 10 sIze. 7.1.1 Equipment for Fluidization Tests When Using Fine Powders

7.1.1.1 Column Size If only small quantities of powder are available, columns down to 2 in (50 mm) in diameter can be used but diameters not less than 4 in (100 mm) are recommended whenever possible. This is to minimize the wall effects and also to increase the accuracy of air flow measurement. As the settled depth should be about 2 ft (60 cm), the column should be about 4 ft tall (120 cm) and preferably increase in diameter above the 4 ft level in order to reduce entrainment. Ideally, there should be two identical columns, one fitted with a high pressure drop distributor for use at low velocities (up to 1 cm/s) when measuring UMF and U MB ' and a second one fitted with a lower pressure drop distributor which will allow measurement of deaeration rates and collapse rates at velocities up to 30 cm/ s. Alternatively, if only one column is available, two different distributors must be used (and replaced as required). 7.1.1.2 Distributors Fig. 34 gives details of a recommended design of a test column. A porous distributor is used throughout, comprising several layers of thick paper glued at the edges with adhesive, and supported by a perforated zinc plate: this arrangement should give a pressure drop across the distributor of about 650 mm WG at a superficial gas velocity of 1 cm/s. Alternatively, a sintered metal (e.g. bronze) plate is used in contact with the powder and the pressure drop increased to the desired value by adding one or more sheets of filter paper under the plate. Distributors of lower resistance can be made be decreasing the number of layers of paper used. Individual distributors may have to be replaced from time to time as they weaken, due to bowing, or become clogged with fine particles. The bed expansion and de-

AERATED FLOW AND HANDLING PROPERTIES

113

_ _ _ _ _ PERSPEX FLUIDIZATION COLUMN

RUBBER GASKETS ~-- _--/_ _

PRESSURE TAPPING - - - _

FIG. 34. Design of a fluidization test column.

aeration rate tests are mostly done at velocities of up to 30 crn/s and the high resistance distributor suitable for measuring UMF and UMB would burst if used at high velocities. Consequently, a different distributor (also made of paper) should be used; this should give a pressure drop about 1/Sth of the other.

7.1.1.3 Leak Testing As the fluidization velocities UMB and UMF are very small for fine powders (less than O.S crn/s), a leak-proof system is essential to give consistent results; a spigotted collar arrangement is used to minimise any side leakage. After assembling, the bed should be treated for leaks in the following way: a steel cover plate is bolted onto the column and the system pressurized to about 900 mm WG. By connecting the pressure tapping in the plenum chamber to a manometer, any leaks present can be detected. Further tightening down on to the bolts or, occasionally, renewing the gaskets, until the manometer level remains constant, usually ensures a leak-proof system. Care should also be taken that the column is vertical and that

114


there are no protruding gaskets because premature bubbling will occur otherwise, giving inconsistent results.

7.1.1.4 Pressure Drop Measurement Bed pressure drops may be measured either with a probe or by means of pressure tappings on the wall near the distributor. The probe may be a stainless steel narrow bore pipe (1/8 in outside diameter, 1/16 in normal bore) with a 1 mm hole 6 mm from the sealed lower end, covered with filter paper to prevent ingress of powder. The probe is connected directly to a water gauge or a mercury manometer. 7.1.1.5 fiir Supply For the de-aeration rate tests, the air supply has to be shut-off suddenly and 2 solenoid valves are used to do this. As one is shut, TO MANOMETER __- - - - - _ , - _ - - , PRESSURE PROBE

ROTAMETERS

GAS -

BED

IXl-'------'-----'

FIG.

~SO-L~E~NO>'D~s~~l~-------,r

35. Air supply for fluidization tests.

the other in the side branch (see Fig. 35) is simultaneously opened so that the gas flow is not interrupted or pressure allowed to build up in the line. It is important to avoid electrostatic effects and control of the air humidity is essential for this. There is a simple but effective, visual method of deciding what the minimum relative humidity


115

should be in order to minimize electrostatics. A copper rod is suspended above the bed, which should be bubbling sufficiently to give some splashing or entrainment. If electrostatic forces are affecting the powder, it will adhere to the earthed copper rod; if the relative humidity is gradually increased, a range of values, usually 30 to 60% will be reached above which little or no particle adherence to the copper rod occurs. 7.1.2 Measurement of Minimum Fluidization Velocity An amount of powder should be charged to the column, sufficient to give a gently-settled bed about 0.6 m deep. The minimum fluidization velocity can then be determined in a standard way from a plot of pressure drop versus superficial gas velocity. Many workers take measurements on reducing the gas flow to zero. More reproducible results are obtained, however, by first allowing the bed to mix well by bubbling freely for several minutes before turning the gas flow rate down to zero. Pressure drop measurements are then taken, increasing the gas flow rate gradually and allowing time for each new velocity to be established throughout the bed. This can be readily verified by watching for the manometer reading to stabilize, and it may require up to 2 minutes at low velocities. Enough readings should be taken in the packed bed region to allow a straight line to be drawn through the data and the origin. The bed depth can also be measured at the same time using a scale fixed to the outside of the column. The minimum fluidization velocity is taken as the intersection of the packed bed and fluidized bed lines as shown in Fig. 36. The overall fluidized bed pressure drop should agree within a few percent with the weight of the powder divided by the cross-sectional area of the bed, provided that the pressure tapping (or the probe) is positioned at or very close to the distributor. The result should be compared with one of the empirical correlations available. Geldart65 found that the Baeyens' correlation gives consistently the most reliable prediction overall for Group A solids (see section 3.2), but even so, anyone value could be ± 60% of the predicted value.

116


,, " ,

1 pressure drop ~p

gas velocity Q / A _

FIG. 36. Minimum fluidization velocity.

A typical equilibrium cracking catalyst (60 microns, Pp = 1500 kg/m3 ) may have a minimum fluidization velocity in the range from 0.1 to 0.2 cm/s. If much higher values are measured, a leak in the air line or below the distributor should be suspected. 7.1.3 Minimum Bubbling Velocity A simple but subjective way of determining the minimum bubbling velocity is by visual observation. By this method, the velocities are noted at which the first distinct bubbles appear when the gas flow is gradually increased and at which bubbling ceases when the gas flow is decreased. The average of the two values may be taken or, alternatively, the values obtained during decreasing flow may be more appropriate. Make sure not to mistake the tiny fissures of "volcanoes" which appear on the surface just below UMB for genuine bubbles; bubbles generally pop in several places on the surface and are about 1 cm in diameter whereas the "volcanoes" usually stay in one place (at a given superficial gas velocity) and can be seen as semi-permanent channels, a few millimeters in diameter. The minimum bubbling velocity can be found less subjectively, according to Geldart66 , when using bed depths greater than about


117

0.6 m, by plotting bed height as a function of gas velocity (see Fig. .........

/

1

bed height

I I

I

I

I

I

i~UMB gas velocity

FIG.

..

37. Minimum bubbling velocity.

37). The intersection of the curves below and above UMB, as shown in Fig. 37, give UMB reproducibly and usually coincides with visual observations. At low bed heights, however, the value determined visually is less than U MB found from the graph. 7.1.4 Bed Expansion The shape of the bed expansion curve can be used as an indicator of the likely behaviour of a group A powder (see section 3.2) and the ratios of bed heights at UMB and UMF are uniquely related to the ratio of the velocities. The measurement of bed expansion is, therefore, a useful check on the velocity ratio. Expansion is easy to measure in the non-bubbling range but once the superficial velocity is above UMB the bed surface fluctuates increasingly. Geldart 66 found a most effective, yet simple technique in using two horizontal rods mounted on a vertical retort stand. The upper rod is sighted against the maximum height attained by the bed at a given velocity, and the lower against the minimum. The two values may then be plotted in the same diagram and the mean value taken. Abrahamsen 67 found the relationship between the bed heights and fluidization velocities in the following form: HMB/HMF = (UMB/UMF)O.22

An alternative method is to use two pressure probes (see earlier) strapped together but with their measuring holes separated by a

118


known distance L in cm (usually 10 cm). A measurement of the differential pressure in cm of water gauge divided by L in cm gives the local bed density. 7.1.5 Bed Collapse, De-aeration Rate The rate of de-aeration of powder under different conditions is an important property in powder handling and processing. The rate of collapse of the powder in a fluidized bed is one way of measuring the rate of de-aeration but it is specific to the operating conditions: the rate depends on, for example, whether or not the plenum chamber (space under the distributor) is vented simultaneously with the stopping of the aeration gas flow. Murfitt and Bransby68 conducted such double drainage tests recently but the most common method is not to vent the plenum chamber; the following description of such method is based on a recently published work of Geldart and Wong28 • Although the primary objective of that work was to identify differences in behaviour between powders in Geldart's groups A and C, analysis of their data and those of other workers has given further insights into the use of the collapse test as a suitable tool for characterising powders. The de-aeration rate measured by this method is particularly relevant, for example, to powder behaviour in standpipes and cyclone diplegs. Since relatively high velocities are to be used, the lower resistance distributor is needed for the test. An experiment to measure the de-aeration rate of a powder is carried out as follows: the gas flow rate is set at a pre-determined level so that the superficial velocity is greater than minimum bubbling velocity and the average bed height HT is recorded. When the gas supply is suddenly turned off by a solenoid valve, the bed begins to collapse and bed height is recorded visually (preferably recorded with a video camera and placing a clock in the field of view) as a function oftime. As the bed height fluctuates considerably before the start of collapse, due to bubbling, up to five repeat tests are made and a numerical average taken which is then plotted. Collapse tests are normally done at a starting fluidization velocity of 10 cm/s.


119

A typical collapse curve for a group A powder is shown in Fig.

FIG. 38. A typical collapse curve for a group A powder.

38. The initial stage of rapid bed collapse is the result of bubbles escaping from the bed. Thereafter the bed surface falls slowly at a constant rate U e , until the bed height approaches He> a height similar to but not necessarily identical with H MF • Although the term "bed collapse" is used, it should not be imagined that the bed deflates by a uniform reduction in voidage throughout its height. Observations of cracking catalyst in a back-lit two-dimensional bed clearly show that the powder settles out continuously in the same way as a sedimenting liquid suspension; there is a clearly visible interface between the powder which has settled and that above which is still fluidized. The interface rises to meet the surface between He and Hs by consolidation as the remaining gas is squeezed out. The basic information derived from a plot of bed height against time is the intercept HD obtained by extrapolating the straight line back to t = 0 and the collapse rate U e calculated from the slope of the line. The time required for group A powders to deaerate during the sedimentation stage is dependent not only on the collapse rate U e but also on the height through which the bed surface has to fall. Geldart and Wong28 used a standardized collapse time defined as the ratio of the collapse time and H MF , and correlated it with the primary properties of the particle-gas system. Cohesive powders in group C display a different collapse rate as shown in Fig. 39. The bed does not bubble even at a gas velocity

120


t

---

bed height

t= 0

time--_

FIG. 39. A typical collapse curve for a group C powder.

of 8 cm/s and bed expansion is caused by the presence of cavities and cracks of various inclinations. When the gas is switched off the largest cracks close up rapidly and further collapse proceeds more slowly as in the final consolidation stage which occurs with group A powders. Because of the channelling, such powders deaerate more quickly in the initial rapid collapse caused by cavingin of cracks. The analysis of the experimental data is based on a first order process which depends on the difference between the depth of the bed H at any time t and the final settled height Hs:

dH/dt = -k (H - Hs) where k is de-aeration rate constant which decreases with both increasing gas viscosity and settled bed depth, and increases with increasing cohesiveness. The above analysis is similar to that used in modelling the compression stage in settling of solids in liquids.

7.2 SIMPLE DE-AERATION TESTS Besides the de-aeration tests in fluidization, there is a very simple test employed in industry to assess aeratability of a bulk solid, and also the rate of de-aeration. A simple procedure is to place a sample of the powder in a glass container, usually a measuring cylinder of 2 litre capacity, with the top closed tightly. The container and the sample is then vigorously shaken, usually by repeated inversion, for one minute


121

or some other specified time. For aerability, the volume increase of the aerated solid in relation to the original, settled volume is noted. The rate of de-aeration can sometimes also be measured in this way, by noting the volume change of the solids with time; this is only possible with solids of extended de-aeration times and the measurement is rather crude even then. It is a quick estimate method, however, and its accuracy may be improved by making a video recording of the settling of the solids. It seems, however, that if such more sophisticated recording of the measurement is to be made, it would warrant the use of a more reproducible experimental set-up than just manual shaking, i.e. the de-aeration test based on fluidization. The above described procedure is only used as a first estimate of powder aerability and, as it is merely a relative measurement, it probably does not need to be standardized other than within a particular organization or industry. For more reproducible and accurate results, the de-aeration test by fluidization described in the previous section is recommended.

7.3 PERMEABILITY The rate of gas flow through a packed bed increases linearly with applied pressure drop. Bed permeability, or the "permeability factor", is the gas flow rate (m 3/s) per unit pressure drop (N/m2), per unit cross-sectional area of the bed (m2) times the bed depth (m), giving the final units of the factor in m4fN" s. The factor as defined above also depends on the gas viscosity (refer to Fig. 4 in section 2.5) and, unless the gas conditions are controlled or specified, it is probably better to express permeability as the gas flow rate at unit viscosity, i.e. in units of m2 • As permeability depends on specific surface area of the powder, it is used for measuring surface area of powders - see section 2.5 for details. Permeability is also highly sensitive to porosity (voidage) ofthe powger, as can be seen from the Carman-Kozeny equation, Fig. 4 section 2.5, and the state of compaction is,

122


therefore, a critical factor controlling permeability. All of this means that permeability is not a useful way to describe powder behaviour unless at precisely specified porosity and, consequently, the use of permeability is very limited in practice. Most of the available permeameters are in fact used for surface area measurement: the methods and instruments are divided into constant flow instruments and variable flow instruments, as described in section 2.5.1.

7.4 DUSTINESS Dustiness (or dustability) is the propensity of fine particles (dust) to be separated from the main bulk of the powder and become airborne. According to British Standard BS 2995, dust is defined as any particulate material finer than 75 microns. The process which causes the dust to become airborne is sometimes referred to as pulvation. Dustiness, or more appropriately, the ways of preventing it, are of increasing importance in handling of powders due to the growing emphasis on health and safety, and also on loss prevention. It is not surprising, therefore, that the interest in testing dustiness is growing and that companies like BORS, Unilever, Monsanto, Michelin and Warren Spring Laboratory, for example, have a wide-ranging programme of dust suppression and abatement. Whether or not a given particle is going to be picked up and become airborne obviously depends on the air flow velocity, but also on the particle weight (hence its size) and the cohesive forces holding the particle in the bulk. Dustiness as a material property is, therefore, a composite property affected by several factors and it is best tested directly. The pulvation process used in such testing, however, has to be standardized and fully reproducible. The methods available for testing dustiness vary in scope from British Standards to those published in the scientific literature. Schofield et al. 69 proposed a fluidization test for dustiness testing. The equipment consists of a 7.62 cm diameter glass tube


123

with a porous base as a distributor. The sample is placed in the tube and is fluidized by air supplied by a positive displacement pump. The dust elutriated from the bed is drawn through an Anderson sampler by another pump. There is an option for recycling the fluidizing air when the powder under test is prone to drying (presumably, provision for cooling of the air has to be made to prevent temperature increases in such are-circulating system). Typically, 0.42 kg of sample is used in the test, under superficial velocities from 0.076 to 0.152 m/s. In order to minimize the effects of the material's stress history, the tube with the sample is taken out, sealed at the top and inverted 10 to 20 times before its replacement into the apparatus. The bed is sampled after the initial 30 seconds of fluidization under a given fluidizing velocity. The sampling period varies with the dustiness of the material so that sufficient amounts are collected on the Anderson sampler glass plates and yet no dust piles form to cause reintrainment. The sampler gives seven points on the particle size distribution of the emitted dust and the results of the test are plotted as grade emission curves, giving cumulative rates of emission in mg/min against particle size. Schofield et al. used the test to study the dustiness of different materials and the effects of fluidizing velocity, moisture content and particle size 69 and made some fairly predictable observations of:

* large differences in dustiness between different materials * increased emissions with increased velocities * large reductions in dustiness with increasing moisture content * reduced dustiness with increasing mean size of the test powder No results on the reproducibility of the method were given. A simple elutriation apparatus for measuring dust in powders is described by Rhoden70 • This provides for passage of air upwards through a sample of powder at a specified rate, for a specified time. The loss of weight of the sample then determines the dust content: the dust content is the total elutriable content, i.e. virtually all of the dust above the cut size determined by the air flowrate, cross-section of the cell, the density of the dust and the

124


viscosity of the gas. The apparatus consisted of air supply, pressure regulator, flowmeter, powder cell and filter, all in series. Rhoden used his apparatus for comparing the effects of de-dusting agents and the dosage: the fines contents did not change between experiments but the presence of the de-dusting agent obviously agglomerated the fines with the coarse particles and reduced their removal from the bed. This method clearly gives an exaggerated figure for dustiness by measuring the total dust contents elutriable. In normal factory handling, only a fraction of the potentially airborne dust content is liberated. Wells and Alexander71 described another method of measurement of dustiness but which measures only the dust which becomes airborne after pouring the powder through a standard height in a cabinet of standard proportions. This is not dissimilar to a patented Procter and Gamble method72 in which a sample is fed from a hopper fitted with a slide, also into a cabinet. Air is drawn through the falling powder at right angles onto a filter pad of a dust sampler operating at flow rates between 450 and 550 litres/minute. A baffle plate prevents all of the powder from being drawn into the sampler. The method by Wells and Alexander uses lower air velocities (the extraction rate is only 50 Htres/minute) and does not, therefore, require a deflector plate. The method does not necessarily release all of the potentially airborne particles in the sample; in fact, the dust release may be constant in many repeated pourings of the same sample. Either total airborne dust samples may be collected on a filter (i.e. all particles that are still airborne on leaving the box, i.e. smaller than about 10 microns) or only the respirable fraction may be collected by extraction through the Hexhlet sampler. The Rolling Drum Dust Generator described in a recent BMHB review54 is a variation on the same idea except that ituses a tumbler similar to those used in attrition tests for the dust generation and a cascade impactor to sample the dust cloud. The drum is cono-cylindrical in shape, consisting of a central cylindrical section (diameter 305 mm and 454 mm long) with coni-


125

cal pieces at each end. There are eight radial lifting bars welded onto the inside of the cylinder. One of the two conical end pieces has a hole through which a nozzle is fitted, sealed with an O-ring, to join the drum to the cascade impactor sampler. The other end is open to atmosphere. The sample used is a Mark II Anderssen IAFCM particle fractionating sampler followed by a filter; alternatively, a shortened Anderssen sampler, a filter only or a light obscuration detector may be used. In the test, the drum is loaded with a 100 g sample and rotated at 30 rpm for one minute. The sampling rate is 28.3 litres/minute and the sampling time once again depends on the dustiness of the material, bearing in mind that the impactor must not be overloaded. This apparatus also measures only the respirable fraction of the dust less than about 9 microns and is capable of determining the dust emission curves or just an overall emission rate of dust finer (or coarser) than 9 micron. Finally, a test according to a British Standard73 for dust in filling materials should be mentioned. This is used to measure the dust content in stuffings for bedding, upholstery or toys. It is based on agitation of a sample of the stuffing (250 to 500 g) by an air stream under controlled conditions and measurement of released particles carried by the air stream through 100 mesh screen (150 microns) and deposited on a 170 mesh screen (100 microns). The test takes place in a cylindrical container 229 mm in diameter and 305 mm high, closed on top by a lid and standing on a base, with at least 38 mm gap between the bottom of the container and the base. The two test sieves are placed at the bottom of the container. Compressed air is blown in through a downward-pointing nozzle near the top of the container and the test lasts one minute. The dust index is the amount deposited on the 170 mesh sieve as percent of the total weight of the sample tested. The test is obviously designed to measure only the coarse dust; a 2-hour period of exposure of the sample to room air is recommended prior to the test. In conclusion to the section on dustiness, the pattern that emerges is much the same as for most other tests described in the guide. There are many different ways of measuring dustiness,

126


depending on the material under test and on the type of process which is being simulated. Either the total elutriable dust may be measured or varying fractions of it, as released in gentle flow, tumbling or air agitation. Some test methods provide for measuring only the respirable range of particle size. The dust emission rate curves give probably the fullest information but the overall emission rate below a certain size is probably sufficient for most purposes. The tests are particularly useful for relative comparisons when evaluating the effects of humidity, dust suppression agents or some process variables used in the material production. All of the tests quoted in this section are affected by humidity of the air and should ideally be carried out in a humidity and temperature controlled environment. Relative humidity of about 10% is said to produce the worst case of dustiness69 • The BORS Working Group on Dustiness Estimation are currently working on harmonization of dust scales and developing and comparing four alternative procedures and apparatus74 with the view of working towards a standard test for dustiness. A review of this development is to be published75 but was not available at the time of writing this guide. 7.5 FLOODABILITY Floodability is the propensity of powders for flooding, i.e. an uncontrolled liquid-like flow out of a hopper. Flooding is attributable to the self-fluidization of a mass of particles caused by the sudden breakage of an arch or by the flow of material into or through a hopper which is wholly or partly empty. Although flooding is clearly associated with aeration and deaeration, de-aeration or fluidization tests alone are not sufficient to establish floodability. A long de-aeration time and large bed expansion observed in a simple de-aeration test usually, but not always76, coincide with a high likelihood of flooding. Bruff6 used another, more reliable, test as an indicator of flooding. A glass tube containing 200 g of the powder is aerated manually by shaking and then raised up from a table thus allowing


127

the powder to flow out. This is done after various time intervals, allowing the powder to settle, and the diameter of the heap formed is measured. The ratio of the maximum to minimum diameters measured in this way gave the best agreement with Bruffs experiences related to flooding. Carr36 proposed a flood ability index which is determined by assessing a number of characteristics such as flowability, dispersibility or angles of repose, fall and difference. A powder is considered more flood able the greater the angle of difference, dispersibility and flowability. Although rather crude, Carr's index is said to be moderately successful in predicting the propensity of a powder for flooding. Recent work by Geldart suggests that the ratio of the minimum bubbling and fluidization velocities, as obtained from fluidization tests (section 7.1), can be a measure of those properties which influence flooding, with the largest values corresponding to the powders which behave worst. According to Geldart28 , all powders in group A (refer to his classification in section 3.2), most in group AC and some in group C are likely to give flooding problems. Within group A, the probability increases as UMa!UMF increases, and within group C the probability decreases as the material becomes progressively more cohesive i.e. finer and softer. Whether a material actually will flood depends also on geometrical, design and operational factors such as whether the bin is a mass flow hopper, the degree of fill, solids charge and discharge rates, whether air injection is used and how much is supplied and where. There might be a way of assessing flood ability indirectly from other powder tests, as Carr and Geldart suggest, but there is still a need for a standard, direct floodability test which could be used as a common yardstick; further development in this area is needed. 7.6 APPARENT VISCOSITY TESTS There have been several developments recently in testing the stress-strain behaviour of aerated powders. Although there is

128


no standard procedure nor equipment yet available, the recent development, at three research centres (Universities of Loughborough and Delft, and Cranfield Institute of Technology) of the Couette-type viscometer adapted for measurement of aerated powders deserves a mention here. The idea of testing the apparent viscosity of powders in different states of aeration is not new: various viscometers have been used before for this such as rotating or vibrating viscometers, tube viscometers, falling spheres or rising bubbles. They all suffered from the disturbance of the powder structure by the measurement technique in one way or another. The Couette-type, concentric cylinder viscometer, preferably with both cylinders rotated independently, offers the possibility of generating a wide range of strain rates at aeration rates ranging from zero to full fluidization, while minimizing the centrifugal settling of the particles within the annulus. Scarlett and HobbeF7 have reported the use of the concentric cylinder viscometer for measurement of stress-strain relationships of fluidized alumina. By rotating both cylinders independently, they were able to create a controlled shear pattern in the annulus. The fluidizing air was introduced through a stationary distributor. The tests showed a distinct asymmetry in the results in that the rotation of the outer cylinder had a much greater effect on the fluidizing behaviour than did the inner cylinder. Thus, there was a significant difference in torque where the same difference in speed was produced by a fast moving inner cylinder and slow moving outer than for the reverse situation. The plots of strain rate against shear rate at various superficial gas velocities provided some valuable insight into particle mobility when fluidized. Lloyd and Webb78 have also designed a similar viscometer but used it at aeration rates less than and up to full fluidization. They experienced rather long time periods (up to an hour) before steady-state conditions were reached in the test bed and had to make modifications to the geometry to prevent preferential passing of the fluidizing air near the surfaces of the two cylinders. Neither particle attrition nor size segregation within the con-


129

tinuously failing bed of powder were checked but the work has provided some insight into powder flooding. Finally, there has also been a similar development of a concentric cylinder viscometer for testing of fluidized grains up to 3 mm in size at Silsoe College of Cranfield Institute of Technology but a full report had not been published by the time of preparation of this guide. In this case, only the outer cylinder rotates and the fluidizing air is introduced through a porous plate at the base of that cylinder. It can be concluded that the Couette-type viscometer may well find its place in standard powder testing because it can measure the stress-strain relationships for aerated powders and thus provide a basis for design of equipment for handling of such powders.

8 Conclusions, Future Work

The multitude of the properties and tests considered in this guide well demonstrates the complexity of bulk powders. The status of the testing of the different powder properties is not the same, though, and there is a considerable imbalance between the various sections and subsections of this guide. In general, when the tests are designed for specific materials and specific applications in scale-up and design then, usually, they are well developed and standalidized, at least in some specific industries. Thus, for example, the shear cell tests and wall friction tests for hopper design; moisture content measurement or the grindability tests in comminution are well catered for. There are notable exceptions, however, like the conveying angles (angle of slide or the angle of rise) where there is little or nothing in terms of recommended test procedures and yet such information seems vital in design of conveying and handling systems. There is a category of tests or powder properties where a need for better understanding has recently been recognized by BMHB, IFPRI, Warren Spring Laboratory or various University departments, and work is under way to fill the gaps. Such subjects include attrition, dustiness, floodability, fluidization, apparent viscosity of aerated powders and morphology of particles and its relationship to handling properties of powders. I leave those areas out of the discussion below. It is in the relatively simply-measured properties where most problems lie. Bulk densities of powders (particularly the tap 131

132


density), angles of repose (particularly the larger scale tests and the conveying angles), compaction test, and the direct methods of measuring internal friction and the failure function require further development of recommended procedures. In understanding, the greatest question to be yet resolved is: what is the best, yet simple, test to be used to monitor the cohesivity or flowability of powders? There are, of course, the direct (or almost direct) ways of measuring it (the cohesion tester or the tensile testers) but neither the tests nor the equipment are really simple. Angles of repose or spatula are simple but, unfortunately, become quite meaningless as the powder gets more cohesive. I personally think that the way powders compact can be directly related to cohesivity; the recent evidence of the usefulness of the Hausner ratio in many different applications points in this direction. The compaction test should also be better defined and more widely used. It is not surprising that the answer to testing of cohesivity may lie in packing properties; the voids between the particles usually occupy a larger volume than the particles themselves and, in many processes involving powders, these voids determine the powder behaviour-. By far the greatest problem I have encountered during preparation of this guide is the general lack of understanding (with a few notable exceptions of institutions and individuals engaged in powder technology research) of the problems involving powders and lack of awareness of the tests and procedures available. Many companies, including some manufacturers of powder handling equipment, use the "suck it and see" method and have expressed a great interest in this guide even before its publication. I think there is an overwhelming need for a manual of powder testing and this guide, together with its companion book when it is published, should satisfy this need. There is also scope for industrial, post-experience courses on testing of bulk powders, dealing not just with the academic niceties of single particle characterization, hopper design, fluidization or comminution but with many of the simpler tests and practical recommendations useful for a nonspecialist. We in Bradford are going to pursue this option too.

9 References

1. T. Allen and A.A. Khan, Chem. Eng., 238 (1970), CE 108-112. 2. P.M. Plowman, Practical aspects of sampling, Bulk Solids Handling, 5(6) (December 1985), 1259-65. 3. British Standard BSlO17, Part 1, 1977, Sampling of coal and coke. 4. P.M. Gy, Sampling of particulate materials-theory and practice, Elsevier, Amsterdam, 1982. 5. J.W. Merks, Mechanical sampling systems for high capacities, Part 1, Bulk Solids Handling, 5(6) (December 1985), 1253-6; Part II, 6(1) (February 1986), 115-19. 6. J.W. Merks, Sampling and weighing of bulk solids, Trans Tech Publications, Clausthal-Zellerfeld, 1985. 7. British Standard BS 4550, Methods for testing cement, Part 1: 1978Sampling. 8. T. Allen, Particle size measurement, 3rd edition, Chapman and Hall, London, 1981. 9. L. Svarovsky, ed., Solid-liquid separation, 2nd edition, Butterworths, London, 1981. 10. J.K. Beddow, ed., Particle characterization in technology, Volume II, Morphological analysis, CRC Press Inc., Boca Raton, 1984. 11. A.R. Abrahamsen and D. Geldart, Behaviour of gas-fluidized beds of fine powders, Part I. Homogeneous expansion, Powder Technology, 26 (1980), 35-46. 12. British Standard BS 4550, Methods for testing cement, Part 3.2 : 1978 Density test. 13. British Standard BS 3483, Methods for testing pigments for paints, Part B8 : 1974 Determination of density relative to water at 4°C. 14. British Standard 812, Methods for sampling and testing mineral aggregates, sands and fillers, Part 2 : 1975 Physical properties. 133

134


15. British Standard 1016, Methods for the analysis and testing of coal and coke, Part 13 : 1980 Test special to coke. 16. British Standard 1377 : 1975, Methods of tests for soils for civil engineering purposes. 17. S. Ergun, Anal. Chem., 23 (1951), 151. 18. D. Geldart, Gas Fluidization, I. Chern. E. Short Course, Bradford, 1986. 19. B. Buczek and D. Geldart, Powder Technology, 45 (1986), 173-6. 20. British Standard BS 4550, Methods of testing cement, Part 3, Section 3.3 : 1978 Fineness test. 21. British Standard BS 4359, Methods for determination of specific surface of powders, Part 2 : 1971 Air permeability method. 22. S. Brimaier, P.H. Emmett and E. Teller, I. Amer. Chem Soc., 60 (1938), 309. 23. A.W. Jenike, Storage and flow of solids, Bulletin No. 123 of the UTAH Engineer Experiment Station, University of Utah, Salt Lake City, 1970 (sixth printing). 24. J.e. Williams, A.H. Birks and D. Bhattacharya, The direct measurement of the failure function of a cohesive powder, Powder Technology, 4 (1970/71), 328-37. 25. J.A. Monick, Measuring the tackiness of detergent powders, Proceedings of Chemical Manufacturers, (May 1966), 108-12. 26. D. Geldart, Types of gas fluidization, Powder Technology, 7 (1973), 285-92. 27. D. Geldart, private communication, 1986. 28. D. Geldart and A.e.Y. Wong, Chem. Eng. Sci., 40 (1985), 653. 29. J.C. Williams, private communication, 1986. 30. J.R.F. Arthur, T. Dunstan and G.G. Enstad, Int. I. Bulk Solids Storage in Silos, 1(2), (1985), 7-10. 31. T. Tamura and H. Haze, Determination of the flow of granular materials in silos, Bulk Solids Handling, 5(3) (June 1985), 633-40. 32. T. Yokoyama, K. Fujii and T. Yokoyama, Powder Technology, 32 (1982),55. 33. J. Boden, PhD Thesis, University of Bradford. 34. J.C. Williams, lifting lid tester, private communication, 1986. 35. P.L. Bransby, Current work in materials handling at Warren Spring Laboratory, The Chemical Engineer, (March 1977), 161-4. 36. R.L. Carr, Evaluating flow properties of solids, Chem. Engng., (January 1965), 163-8. 37. British Standards 4140 (also ISO 2927 -1973), Part 9 : 1970, Determination of angle of repose. 38. H.N. Wilkinson, C.H. Duffell, A.R. Reed and J. Bunting, Bulk

REFERENCES

135

Solids Physical Property Guide, British Materials Handling Board, Ascot, 1983. 39. D.A. Augenstein and R. Hogg, An experimental study of the flow of dry powders over inclined surfaces, Powder Technology, 19 (1978), 205-15. 40. Book of ASTM Standards, Part 9, American Society for Testing and Materials, Philadelphia, 1978, 45. 41. D.A. Hall and 1.G. Cutress, The effect of fines content, moisture and added oil on the handling of small coal, J. Inst. Fuel, 33 (1960), 63-72. 42. D. Geldart, N. Hamby and A.C.Y. Wong, Powder Technology, 37 (1984) 25. 43. Metallic powders-determination of apparent density, ISO 3923 Part 1 : Funnel method. 44. Metallic powders-determination of apparent density, ISO 3923/2 1981 (E), Part 2 : Scott volumeter method. 45. Metallic powders-determination of apparent density, ISO 3923 Part 3 : Oscillating funnel method. Note: This standard has just been reviewed by the Sub Committee ISM/65/2 and will be published shortly. The summary given below is based on the document TC119 N 542 (85/35035) of BSI. 46. 1967 book of ASTM standards, Part 7, ASTM, March 1967,231-8. 47. M.e. Kostelnik and 1.K. Beddow, New techniques for tap density, in Modern developments in powder metallurgy, Vol. 4. Processes (ed. H.H. Hausner), Plenum Press, New York, 1970,29-48. 48. ASTM D 4164 - 82, Mechanically tapped apparent packing density of formed catalyst particles. 49. ASTM D 4180 - 82, Vibrated apparent density of formed catalyst particles. 50. H.H. Hausner, Friction conditions in a mass of metal powder, Int. J. Powder Metallurgy, 3(4) (1967), 7-13. 51. A. Adler, Flow properties of metal powders, Int. Journal of Powder Metallurgy, 5(1), (1969). 52. M.C. Kostelnik, F.H. Kludt and 1.K. Beddow, The initial stage of compaction of metal powders in a die, Int. Journal of Powder Metallurgy, 4(4), (1968). 53. T.A. Roberts and 1.K. Beddow, Some effects of particle shape and size upon blinding during sieving, Powder Technology, 2 (1968), 121-4. 54. BMHB, Particle Attrition State of the Art Review, TransTech Publications, Clausthal-Zellerfield, 1987. 55. Kirk-Othmer concise encylopedia of chemical technology, 10hn Wiley & Sons, New York, 1985.

136


56. British Standard BS 1016 Part 20 : 1981-Determination of Hardgrove grindability index of hard coal. 57. V.c. Marshall, ed., Comminution, Institution of Chemical Engineers, 1975. 58. F.e. Bond, Crushing and grinding calculations, Part 1, Brit. Chem. Eng., 6(6), (July 1961), 378-85. 59. British Standard BS 812, Methods for sampling and testing of mineral aggregates, sands and fillers, Part 3 : 1975 Mechanical properties. 60. Book No. 550 -1970, Classification and definitions of bulk materials, Conveyor Equipment Manufacturers Association, Washington, 1970. 61. H.F. Yancey, M.R. Geer and J.D. Price, An investigation of the abrasiveness of coal and its associated impurities, Trans. Amer. Inst. Min. Metall. Engrs. 190 (March 1951), 252-68. . 62. British Standard BS 1016, Methods for the analysis and testing of coal and coke, Part 19 : 1980 Determination of the index of abrasion of coal. 63. British Standard BS 6212 : 1981 (ISO 3271 - 1975), Method for determination of tumbler strength of iron ores. 64. ASTM D 4058 - 81, Attrition and abrasion of catalysts and catalyst carriers. 65. D. Geldart and J. Baeyens, Proc. Int. Symp. Fluidization and its Applications, Toulouse (1973), 263. 66. D. Geldart, private communication, 1986. 67. A.R. Abrahamsen, MSc Thesis, University of Bradford, 1980. 68. P.G. Murfitt and P.L. Bransby, Powder Technology, 27 (1980),149. 69. C. Schofield, H.M. Sutton and K.A.N. Waters, The generation of dust by materials handling operations, Conf. paper. 70. F. Rhoden, Apparatus to measure dust in powders, Lab. Pract., 24 (1975), 247. 71. A.B. Wells and D.J. Alexander, A method for estimating the dust yield in powders, Powder Technology, 19 (1978), 271-7. 72. Procter and Gamble Ltd., Method and apparatus for measuring dust properties of granular materials, Br. pat. 1343 963 (1974). 73. British Standard 3400 : 1967 Methods of test for dust in filling materials. 74. BOHS Technical Guide No.4 on Dustiness estimation. 75. K.N. Davies, C.M. Hammond, R.W. Higman and A.B. Wells, Progress in dustiness estimation, BOHS Conference, 1986. 76. W. Bruff, J. of Engng for Ind. (ASME), 91 (1969),323. 77. B. Scarlett and E.F. Hobbel, Bergen, Stress-strain behaviour of aerated powders, EFCE publication series No. 49, CHR. Michelsen Inst., 1985, 4fr61.

REFERENCES

137

78. P.J. Lloyd and P.J. Webb, The characterization of the flow of aerated powders, 1. World Congress Particle Technology, Part IV, NMA Nuremberg, 16-18 April 1986, 357. 79. D. Geldart, Development of standardized tests for characterizing particulate systems in motion, A discussion document, Bradford, 29 May 1984. 80. N. Hamby and I. Jackson, The effect of moisture on powder mixing processes,4th European Conf on Mixing, 27-29 April, 1982, BHRA, Cranfield, Paper E2. 81. British Standard BS 1016 Part 21 : 1981, Analysis and testing of coal and coke. Determination of moisture-holding capacity of coal. 82. P. Field, Dust explosions, Elsevier, Amsterdam, 1982. 83. K.S.W. Sing, Characterisation of powder surfaces (eds G.D. Parfitt and K.S.W. Sing), Academic Press, 1976. 84. BMHB Draft Code of Practice for the design of silos, bins, bunkers and hoppers, March 1985.

Index

Abrasion tests, 105-6 Abrasiveness index, 105 Aerated powders bulk density, 82, 83-4 classification of, 38-40 fluidization properties, 111-20 viscosity of, 127-9 Aggregates particle density of, 21 shatter resistance of, 104 Air pyknometry, 17-20 Ajax Equipment attrition test equipment, 107 Wall Friction Tester, 47-8 WSL Cohesion Tester, 68-71 equipment described, 68, 69 limitations of, 71 purpose of, 68 test procedure for, 68, 70-1 WSL Tensile Test Machine, 63-5 equipment described, 64 test procedure, 65 Alumina angle of repose, 72-3 fluidized, viscosity of, 128 Angle of internal friction definitions of, 42, 48, 54 direct measurement of, 44, 53-4 indirect determination of, 48-53 use of, 42-3 see also Effective angle of internal friction repose definitions of, 71-2 measurement of, 72-4 rise, 75--6

Angle of-contd. slide, 75 sliding, 76 spatula, 76-7 test procedures required for, 77 wall friction, 47-8 Apparent density, test procedures for, 84-9 Apparent viscosity tests, 127-9 ASTM standards flowability, 77 friability, 107, 108 sampling procedures, 6, 7-8 tap density, 89, 90, 91 Attrition characteristics relating to, 97-109 definition of, 106 Australian Standards, sampling procedures, 6

Ball Mill Index, 102-3 Ball mills, 100, 101-2 Bed collapse (of fluidized beds), 118-20 expansion (of fluidized beds), 11718 pressure drop method (for particle density), 22 voidage method (for particle density), 21-2 Bias testing, mechanical sampling procedures,.8 Bond ball mill grindability test, 100, 101 Work Index, 101-2 139

140

INDEX

Catalysts-contd. Bone char/particles envelope density of, 23 minimum fluidization velocity for, vibration testing of, 104-5 116 British Occupational Hygiene Society particle density of, 20 (BOHS), Working Group on tap density of, 91 Dustiness Estimation, 126 vibration testing of, 104 British Standards see also Group A powders Characteristic particle size, 12-13 abrasion tests, 105 Circularity (shape factor), 15 aggregates/sands/fillers, 21, 104 alumina, 72 Classification angle of repose, 72-4 aerated powders, 38-40 cement sampling, 9 de-aerated powders, 35-7 shear cell testing approach, 36-7 coal/coke, 105, 107, 108 compacted bulk density, 93 tackiness approach, 37 compaction testing, 94-5 pneumatically conveyed powders, compression testing, 104 40 dustiness of filling materials, 125 Coal abrasiveness of, 105 friability, 107, 108 compression testing of, 104 iron ore, 107 manual sampling from trucks, 9moisture content, 30, 31-2 particle density, 17, 21 10 sampling procedures, 5, 9 mechanical sampling of, 6 soils, 17,31,60,93,94-5 moisture content of, 32, 33 strength properties, 104 shatter resistance tests, 104 Brunauer-Emmett-Teller (BET) Coarse materials failure function of, 61 equation, 28-9 Bubbling (of fluidized beds), 116-17 friability of, 109 see also Group D powders Bucket-type (sampling) cutters, 7 Bulk density Cohesion assessed by tensile strength aerated, 82, 83-4 measurements, 67 compacted, 93 compressibility calculated from, 93 definitions of, 46, 67 definition of, 81 testers, 67-71 Hausner ratio calculated from, 92- Cohesive powders, definitions of, 38, 3 39,40 poured, 84-9 see also Group C powders tap density, 82, 89-92 Comminution factors affecting, 97, 98 research into, 97, 98 see also Abrasion tests; Attrition; Caking end point method, 20-1 Friability; Grindability Carman-Kozeny equation, 22, 25, 121 Compacted bulk density, 93 Carr's Index (for flood ability/ Compaction tests, 93-5 flowability), 78, 127 Compressibility bulk density, 93 Casagrande apparatus, 32 Catalysts uniaxial compression, 94 Compressible powders, particle classification of, 38, 39 density of, 22 friability of, 108

141

INDEX Compression Tackiness Tester classification using, 37 comments on, 60-1 equipment, 59 experimental procedure, 58-60 length-to-diameter ratio effects, 61 storage conditions simulated by,

37,60 Compression tests, friability assessed by, 104 Consolidometer, 94 Conveying angle, 75-6 Couette-type viscometers, 128, 129 Coulter Counter equivalent volume diameter, 15, 16 Cross-belt type (sampling) cutter, 7 Cyclones, attrition in, 106

Dustiness-contd. measurement of-contd. British Standards filling materials method, 125 Procter and Gamble method,

124 Rhodes elutriation method, 1234 Rolling Drum Dust Generator,

124-5 Schofield fluidization method,

122-3 Wells-Alexander methods, 124 Effective angle of internal friction,

39,43-4 Equivalent volume diameter, 15, 16

De-aerated powders classification of, 35-7 flow properties of, 41-66 handling properties of, 67-78 De-aeration rates, 3&-40, 120-1 fluidized beds, 118-20 Density bulk aerated bulk density, 82, 83-4 compacted bulk density, 93 compressibility determined from,

93 definition of, 81 Hausner ratio determined from,

92-3 poured bulk density, 84-9 tap density, 82, 89-92 particle definitions of, 16 measurement of, 17-23 Diverter (sampling) cutters, 6, 7 Drained angle of repose, 72, 74-5 Drop shatter tests, 104 Durham Vibrating Cone apparatus, 77 Dustiness definition of, 122 measurement of, 122-6

Failure function classification on basis of, 36 definition of, 44 large-scale uniaxial test, 61-2 measurement of, 54-62 uniaxial compression methods,

55-8 use of, 45 properties, definitions of, 41-7 Fertilizers, vibration testing of, 104-5 Fillers particle density of, 21 shatter resistance of, 104 Fisher Subsieve Sizer, 27 F100dability definition, 126 determination methods, 126-7 Flow moisture point, 32 Flowability definition of, 77 Hausner ratio for, 92 measurement of, 77-8 particle size limitations, 35 Flowrate tests, 77-8 Fluidization characteristics, 92 classification by, 3&-40

142

INDEX

Fluidization-contd. tests air supply for, 114-15 bed collapse characteristics, 11820 bed expansion characteristics, 117-18 column size used, 112 distributors used, 112-13 electrostatic effects, 115 equipment used, 112-15 leak testing for, 113-14 minimum bubbling velocity determination, 116-17 minimum fluidization velocity determination, 115-16 pressure drop measurements, 114 Fluidized beds attrition in, 106 bed collapse of, 118-20 bed expansion of, 117-18 minimum bubbling velocity for, 116-17 minimum fluidization velocity for, 115-16 Free-flowing particles failure function classification of, 36 size range of, 35 Friability compression test for, 104 definition of, 106 tests, 107-9 vibration tests for, 104-5 Friction. See Internal ... ; Wall friction Funnel method (for bulk density), 85-6 Future developments, 132

Gas adsorption methods particle size used, 29 surface area measured by, 28-9 properties, effects of, 2

Geldart's classification, 38-40 de-aeration collapse characteristics, 39, 119-20 density-size ranges, 38, 39 fluidization properties, 38 tapped-to-aerated densities ratios, 38 see also Group A ... ; ... B. .. ; ...c ... ; ... D powders Gooden and Smith method (for surface area determination), 27 Grindability definitions of, 103 tests, 99-103 Grooved plate method (for internal friction), 44, 53-4 angle of internal friction defined for, 54 Group A powders de-aeration collapse of, 39, 119 density-size range for, 38, 39 fluidization properties of, 38 tapped-to-aerated densities ratio, 38 Group AC powders de-aeration collapse of, 39 density-size range for, 38, 39 tapped: aerated densities ratio, 38 Group B powders de-aeration collapse of, 39 density-size range for, 39 fluidization properties of, 38 Group C powders de-aeration collapse of, 39, 119-20 density-size range for, 38, 39 fluidization properties of, 38 tapped: aerated densities ratio, 38 Group D powders density-size range for, 39 fluidization properties of, 38

Hall test cup, 90 Handling characteristics aerated powders, 38-40, 111-29 de-aerated powders, 35-7, 67-78

INDEX Handling characteristics-contd. pneumatically conveyed powders, 40 Hardgrove index (of grindability), 100, 101 test (for grindability), 100-1 Hardness, 98--9 Hausner ratio, 78, 92-3 Hopper design de-aeration rates in, 38--40 dead spaces estimated, 75 failure properties used, 42-3, 45 floodability effects 126 Hoppers, sampling from, 10 Hosokawa Micron Cohetester, 65 Powder Characteristics Tester, 78, 90,92 IMCO Standard Flow Table Test, 32 Impact tests, 98, 103-4 Internal friction angle of definitions of, 42, 48, 54 direct measurement of, 44, 53-4 indirect determination of, 43-4, 49-53 use of, 42-3 effective angle of, 39, 43-4, 49 definition of, 39, 43-4 Introduction (to this Guide), 1-2 ISO International Standards bulk density methods, 84-9 flowability, 77 Jenike classification scheme, 36 shear cell, 49-50 coarse particles removed before use, 35 Ladle (for sampling), 9 Lea and Nurse apparatus (for surface area), 25-6 equation for, 26

143

Lifting-lid testers (for tensile strength), 66 Liquid limit (of soils), determination of, 31-2 pyknometry, 17, 18 Lloyd-Webb viscometer, 128

Manual sampling, stopped-belt sampling used, 9 Mechanical sampling, 5-8 bias testing of, 8 cutters design requirements for, 5-6 types of, 6-7 Metal powders bulk density of, 84-9 flowability of, 77, 92 MICUM test (for friability), 107, 108 Minerals moisture content of, 32 see also Aggregates Minimum bubbling velocity, 116-17 Minimum fluidization velocity, 11516 Mohr's circles, 43, 44, 52 Mohs' hardness scale, 99, 100 Moisture content compression tackiness affected by, 60 determination of, 29-33 flow moisture point method, 32 gas, effect of, 2, 115, 126 liquid limit methods, 31-2 speedy method for determination, 31 holding capacity, determination of, 33

Oedometers,94 Ores, friability of, 107, 108 Oscillating funnel method (for bulk density), 88--9

144

INDEX

Packed bed porosity, 79-80 measurement of, 81 Particle density apparent, 16 effective (or aerodynamic), 16 comment on, 23 measurement of, 20-3 true, 16 shape, 14-16 size, 12-13 distribution, 13 Pellets compression testing of, 104 friability of, 108 Penetrometers, moisture content testing, 31-2 Permeability, packed bed, 121-2 Permeametry methods, 24-7 ISO recommendations on, 27 particle sizes used, 27 porosity determined by, 81 Perschl Shear Tester, 51-2 Pneumatically conveyed powders, classification of, 40 Porosity compared with void ratio, 80 definitions of, 79-80 measurement of, 81 packed bed, 79-80 void ratio, and, 80 Poured angle of repose, 72 Poured bulk density, 84-9 funnel method for determination of,85-6 oscillating funnel method for, 88-9 Scott volumeter method for, 86-8 test procedures for, 84-9 Primary properties characteristic particle size, 12-13 density, 16-23 imbalance in emphasis on, 1-2 lack of correlation with 'technological' properties, 2, 11, 12 mean size, 13 moisture content, 29-33 particle size distribution, 13

Primary properties-contd. shape factors, 14-16 strength, 98 surface area, 23-9 Principal consolidation stress, definition of, 45 Procter and Gamble (dustiness) test method, 124 Pyknometry, 17-20 Relative density, 17 Relative humidity, effects on dustiness, 126 electrostatic effects, 115 handling properties, 2 Repose angle of definitions of, 71-2 measurement of, 72---4 drained angle of, 72, 74-5 poured angle of, 72 Rhodes (dustiness) test method, 1234

Ring-ball/ring-roll pulverizers, 100, 101 Rittinger number, 101 Ro-Tap method (for tap density), 90-2 Rolling Drum Dust Generator, 124-5 Rotating hammer samples, 7 Sampling general comments, 3-5 'golden' rules of, 3-4 ladle used, 9 manual sampling procedures, 8-10 mechanical sampling procedures, 5-8 primary samples, 5-7 probes used, 10 purpose of, 3 requirements for, 3 secondary sample dividers, 7-8 splitting techniques, 4-5 suction sampler used, 10 'thieves' used, 9

INDEX

Sand displacement method (for particle density), 23 particle density of, 21 shatter resistance of, 104 see also Group B powders Scarlett-Hobbel viscometer, 128 Schofield (dustiness) test method, 122-3 Scope (of this Guide), vii Scott volumeter method (for bulk density), 86-8 Secondary sample dividers, 7-8 Shape factors circularity, 15 handling characteristics affectcd by, 14 Hausner ratio sensitive to, 92 ratio coefficients, 15-16 sphericity, 14--15 surface shape coefficient, 15 surface-volume shape coefficient, 15 volume shape coefficient, 15 Shatter resistance tests, 104 Shear cell tests, classification on basis of,36-7 Shear cells annular shear cells, 50-1 biaxial shear testers, 52 lenike shear cell, 49-50 ring cells, 51-2 rotational shear boxes, 50-2 translational shear box, 49-50 triaxial shear testers, 52-3 Ship cargoes, moisture content of, 32 Silsoe College viscometer, 129 Single-particle properties, 11-33 characteristic particle size, 12-13 density, 16-23 imbalance in emphasis on, 1-2 lack of correlation with technological properties, 2, 11,12 mean size, 13 moisture content, 29-33 particie size distribution, 13 shape factors, 14--16

145

Single-particle properties-contd. strength, 98 surface area, 23-9 Sinters, friability of, 108 Size properties, 12-13 Slide, angle of, 75 Sliding, angle of, 76 Soils angle of internal friction of, 52 bulk density of, 93 compression tackiness of, 60 consolidation testing of, 94--5 moisture content of, 30, 31-2 Spatula, angle of, 76-7 Specific gravity, 17 Specificity (of some tests), 131 Sphericity, 14--15 Spinning riffter, 4--5 Split cell testers (for tensile strength), 63-5 Stockpiles, sampling from, 10 Stokes' diameter, 12 Stopped-belt sampling, 9 Surface area importance of, 23 measurement of gas adsorption methods, 28-9 permeametry methods, 24--7 Surface shape coefficient, 15 Surface-volume diameter, 12-13 shape coefficient, 15 Tackiness classification on basis of, 37 measurement of, 37, 58-61 Tap density definition of, 89 measurement of, 82, 89-92 Tap-Pak Volumeter, 91 Tensile strength definition, 62 measurement methods, 63-6 lifting-lid testers, 66 split cell testers, 63-5 use of, 62-3 Test methods, categorization of, 1

146

INDEX

Transportable moisture limit, 32 Trucks, manual sampling from, 9-10 Tumbler tests (for friability), 107, 108-9 Tyler Ro-Tap device, 91, 92

Voidage compared with void ratio, 80 definition of, 80 Volume shape coefficient, 15

Ultimate tensile strength, definition, 46 Unconfined yield strength, definition, 45 Understanding, lack of in powder problems, 132 United States Bureau of Mines (USBM) Ball Mill Test, 102 coal reference samples, 100 standards. See ASTM standards

Wall friction angle of, 47-8 tester, 47-8 Warren Spring Laboratory (WSL) cohesion tester, 68-71 grindability test, 101 tensile test machine, 63-5 Water content. See Moisture content Wells-Alexander (dustiness) test method, 124 Williams uniaxial compression method (for failure function), 55-8 comments on, 58 equipment used, 55-6, 57 length-to-diameter ratio effects, 56,57

Vibration tests, 104-5 Viscometers Couette-type viscometers, 128, 129 Lloyd-Webb viscometer, 128 Scarlett-Robbel viscometer, 128 Silsoe College viscometer, 129 Void ratio definition of, 80 porosity, and, 80

Yancey-Greer-Price (abrasiveness) test, 105-6

Powder Testing Guide: Methods of measuring the physical properties of bulk powders

Powder Testing Guide: Methods of measuring the physical properties of bulk powders

Powder Testing Guide.. Methods of measuring the physical properties of bulk powders

Handbook of polymer testing - physical methods

Physical Testing of Textiles

Physical Testing of Textiles

Physical Testing of Rubber

Physical Properties of Macromolecules

Physical Properties of Quasicrystals

Physical Properties of Lipids

Physical Properties of Nanosystems

Physical Properties of Polymers

Guidelines for Safe Handling of Powders and Bulk Solids

Physical properties of crystals

Physical Properties of Textile Fibres

Physical Properties of Chemical Compounds

Physical Properties of Liquid Crystals

Food Powders: Physical Properties, Processing, and Functionality (Food Engineering Series)

Guidelines for Safe Handling of Powders and Bulk Solids

Physical properties of carbon nanotubes

Physical Properties of Carbon Nanotubes

The Physical Properties of Organic Monolayers

Handbook of Physical Testing of Paper

Methods of experimental physics, - Polymers. part C Physical properties

Methods of testing concrete, method 3: determination of properties related to the consistency of concrete

Methods of testing concrete, method 3: determination of properties related to the consistency of concrete

Methods of testing concrete, method 3: determination of properties related to the consistency of concrete

Powders and Bulk Solids: Behavior, Characterization, Storage and Flow

Methods of testing concrete, method 3: determination of properties related to the consistency of concrete

Cascade Separation of Powders

Physical Properties of Hydrocarbons vol.1

Powder Testing Guide: Methods of measuring the physical properties of bulk powders