Mixing in Single Screw Extrusion

Citation preview

Mixing in Single Screw Extrusion

Martin Gale

Mixing in Single Screw Extrusion

Martin Gale

iSmithers – A Smithers Group Company Shawbury, Shrewsbury, Shropshire, SY4 4NR, United Kingdom Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118 http://www.rapra.net

First Published in 2009 by

iSmithers Shawbury, Shrewsbury, Shropshire, SY4 4NR, UK

©2009, Smithers Rapra

All rights reserved. Except as permitted under current legislation no part of this publication may be photocopied, reproduced or distributed in any form or by any means or stored in a database or retrieval system, without the prior permission from the copyright holder.

A catalogue record for this book is available from the British Library.

Every effort has been made to contact copyright holders of any material reproduced within the text and the authors and publishers apologise if any have been overlooked.

ISBN (Hard-backed): 978-1-84735-130-2 ISBN (Soft-backed): 978-1-84735-129-6 e-book: 978-1-84735-131-9

Typeset by iSmithers Indexed by Damco Solutions, Hayes, Middlesex, UK Printed and bound by Lightning Source

C

ontents

Preface....................................................................................................................1 1

The Need for Good Mixing in Single Screw Extrusion.....................................3 1.2 Examples of Mixing Problems................................................................9 1.2.1 Polyethylene Pipes and Cables....................................................9 1.2.2 Blow Moulded Bottles..............................................................11 1.2.3 Chalk Filled Polypropylene Pipe...............................................12 1.2.4 Blown Film..............................................................................13 1.2.5 Industrial Blow Mouldings.......................................................13 1.2.6 Production Scrap Re-use..........................................................13 1.2.7 Agglomerates and Gels in Thin Extrusions...............................14 1.2.8 Transparent Polycaprolactone/SAN Blends..............................15 1.2.9 Decorative Wood Grain Effects................................................15 References .....................................................................................................16

2

Dispersive and Distributive Mixing................................................................17 2.1 Definitions and Illustrations..................................................................17 2.2 Dispersive Mixing.................................................................................20 2.2.1 Dispersive Mixing Mechanisms ..............................................20 2.2.2 Dispersive Mixing of Additive Powders Such as Pigments........24 2.3 Distributive Mixing..............................................................................29 2.3.1 Laminar Shear Flow Mixing....................................................29 2.3.2 Measurement of Distributive Mixing Achieved by Laminar Shearing.....................................................................32 i

Mixing in Single Screw Extrusion 2.3.3 Limitations of Lamina Flow Mixing .......................................34 2.3.4 Eliminating Laminar Striations................................................36 References......................................................................................................55 3

Measurement of Mixing.................................................................................59 3.1 The Need for Measurement of Mixing.................................................59 3.2 Striation Thickness Measurement.........................................................60 3.3 Agglomerate Measurement...................................................................61 3.3.1 Microscopy Examination of Thin Samples...............................61 3.3.2 Agglomerate Count for Blown Film.........................................62 3.3.3 Screen Pack Filtration Test.......................................................62 3.4 Influences of Mixing on Product Properties..........................................68 3.5 Preparation of Thin Sections for Optical Microscopy Assessment........69 References......................................................................................................69

4

Single Screw Extruder Stages: Effects on Mixing............................................71 References......................................................................................................75

5

Pellet Handling: A Source of Variable Composition.......................................77 5.1 Introduction.........................................................................................77 5.2 Hopper Design.....................................................................................78 5.2.1 Mass Flow Hopper .................................................................78 5.2.2 Non-mass Flow Hopper . ........................................................79 5.2.3 Round Hoppers ......................................................................80 5.2.4 Square and Rectangular Hoppers . ..........................................81 5.2.5 Ledges and Corners ................................................................81 5.3 Composition Variations........................................................................82 5.3.1 Example 1................................................................................82 5.3.2 Example 2................................................................................82 5.3.3 Example 3................................................................................82

ii

Contents 5.3.4 Other Systems..........................................................................83 5.4 Measurement of Particulate Properties.................................................84 5.4.1 Hopper Flow Tests...................................................................84 References......................................................................................................85 6

Solids Conveying in the Feed/Transport Zone................................................87 6.1 Smooth Feed Zones..............................................................................87 6.2 Grooved Feed Zones.............................................................................90 6.3 Particulate Friction Measurements........................................................96 6.4 Friction in the Feed Zone......................................................................99 References....................................................................................................100

7

Melting.........................................................................................................101 7.1 Melting Mechanism............................................................................101 7.2 Variations in Melting Rate..................................................................103 7.3 Solids Bed Break-up............................................................................105 7.4 Melting Devices..................................................................................107 7.5 Barrier Flight Melting Screws.............................................................115 7.5.1 The Barrier Screw Concept....................................................115 7.5.2 Maillefer Barrier Screw..........................................................117 7.5.3 North American Barrier Screws.............................................118 7.5.4 Combined Barrier Screws and Grooved Feed Zones...............123 7.5.5 Barrier Screw Developments..................................................124 7.6 Other Melting Screws.........................................................................125 7.6.1 Double Wave Screw ..............................................................125 7.6.2 Barr Energy Transfer Screws .................................................126 7.6.3 Stratablend Mixing Screw......................................................126 7.6.4 Shear-Ring Screw...................................................................127 iii

Mixing in Single Screw Extrusion 7.7 Barrier Flight Screws versus Conventional Screws..............................127 References....................................................................................................131 8

Screw Channel Mixing and the Application of Mixing Sections...................135 8.1 Striations: Their Formation and Mixing in the Screw Channel...........135 8.2 Mixing During Melting......................................................................137 8.3 Mixing in the Melt Filled Screw Channel . .........................................137 8.4 Residence Time Distribution (RTD)....................................................144 8.4.1 Concentration Smoothing......................................................147 8.4.2 Variation of Residence Time with Channel Position...............147 8.4.3 Implications of Pressure/Drag Flow Effects............................147 8.5 Mixing Sections..................................................................................148 8.5.1 Maddock Mixer.....................................................................148 8.5.2 Pins and Slots.........................................................................149 8.5.3 Mixer Evaluation Using an Independent Drive.......................152 References....................................................................................................164

9

Interacting Rotor/Stator Mixers...................................................................167 9.1 Overview............................................................................................167 9.2 Turbine Mixing Heads........................................................................168 9.2.1 Stanley (ICI) Mixer................................................................168 9.2.2 Other Turbine Mixers............................................................170 9.3 Woodroffe Key Slot Mixers................................................................171 9.3.1 Gerber (Metal Box) Mixer ....................................................171 9.3.2 Renk (Barmag) Mixer ...........................................................172 9.4 Rounded Cavity Mixers......................................................................176 9.4.1 Rapra Cavity Transfer Mixer.................................................176 9.4.2 Reifenhauser Staromix ..........................................................184 References....................................................................................................186

iv

Contents 10 Floating Ring Mixing Devices......................................................................189 10.1 Introduction.......................................................................................189 10.2 Injection Moulding Check-ring Mixers...............................................189 10.3 Adaption of the Check Ring Mixer to Extrusion................................193 References....................................................................................................196 11 Static (or Motionless) Mixers.......................................................................197 11.1 Mixing Mechanism.............................................................................197 11.2 Static Mixers Used in plastics extrusion . ...........................................197 11.2.1 Helical Mixers.......................................................................198 11.2.2 Honeycomb Mixers...............................................................199 11.3 Application in Heat Exchangers ........................................................200 11.4 Disadvantages.....................................................................................200 References....................................................................................................202 12 Incorporation of Liquid Additives and Dispersions by Direct Addition........203 12.1 Viscosity Differences...........................................................................204 12.2 Incorporating Liquid Additives . ........................................................204 12.3 Some Examples of Liquid Injection Processes.....................................208 12.3.1 Polybutene in Pallet-wrap and Silage-wrap Film....................208 12.3.2 Injection of Liquid Colours (General)....................................208 12.3.3 Wire Insulation Colouring.....................................................209 12.3.4 Fibre Extrusion......................................................................213 12.3.5 Skin Colouring Pipes and Profiles...........................................214 12.3.6 Crosslinking Polyethylenes.....................................................216 12.3.7 Silicone Lubricant Injection ..................................................220 12.3.8 Extrusion Foaming.................................................................220 References....................................................................................................227 v

Mixing in Single Screw Extrusion 13 Dispersive Mixing of Fillers and Pigments....................................................229 13.1 Formation of Agglomerates...............................................................229 13.2 Formation of Filler Agglomerates in a Single Screw Extruder............230 13.3 Starved Feeding to Avoid Agglomerate Formation.............................234 13.4 Dispersive Mixing Using Polymer Powders........................................239 13.5 Dispersive Mixing Using Polymeric Waxes........................................239 References....................................................................................................242 14 Dispersive Mixing Applied to Polymer Blending...........................................243 14.1 Polymer Blends..................................................................................243 14.2 Polymer Scrap....................................................................................246 14.3 Polymer Waste...................................................................................246 14.4 Blending Immiscible Viscous Fluids...................................................246 14.5 Polymer Blending Mechanisms in a Single Screw Extruder................250 14.6 Break-up of Fibrils into Droplets.......................................................252 14.7 Polymer Blending in Single Screw Extrusion: Overall Mechanism.....254 14.8 Mixing by Controlled Continuous Chaotic Advection.......................257 14.9 Blending Mixed Polymer Waste: Comparison of Twin Screw and Single Screw Extruders................................................................259 14.10 Elongational Flow Mixing.................................................................261 14.11 Elimination of Gels............................................................................262 References....................................................................................................263 15 Compounding with Single Screw Extruders..................................................269 References....................................................................................................270 Appendix – Preparation of Microtome Sections for Assessment of Dispersive and Distributive Mixing..............................................................273 Flattening Sections........................................................................................273 vi

Contents Trimming the Block......................................................................................274 Flattening the Rolled Sections.......................................................................276 Holey Sections..............................................................................................276 Brushing Flat................................................................................................276 Distortion.....................................................................................................276 Washing and Mounting................................................................................277 Abbreviations......................................................................................................279 Index...................................................................................................................281

vii

Mixing in Single Screw Extrusion

viii

P

reface

Most extruded plastics products contain additives and therefore mixing is involved at some stage in their production. Mixing is normally associated with twin screw extruders, and conversion to products associated with single screw extruders. Consequently, the latter’s potential mixing performance and economic gains tend to be overlooked. During the many years I was involved with the Smithers Rapra training course: Exploring Extrusion, the subject of mixing in single screw extruders always generated a lot of interest. It seemed, therefore, logical to treat this subject in more detail, particularly with regard to present day economics. The attendees of these training courses came from a very wide spectrum of expertise and experience. These included engineers, chemists, supervisors, plant operators, quality controllers, technical service and sales people. I decided to write this book with this readership in mind. As a consequence, I have paid little attention to mathematical derivations and instead concentrated on the results. In any case, extrusion theory is very well covered by a number of books on extrusion to which I have referred. Most of these books, which cover specific topics in depth have individual authors for each chapter, each one an expert in their field. By writing a book completely on one’s own, this advantage is denied. On the other hand, it gives the author complete freedom to decide what to include and what to omit, to link the chapters together and to make them as long or short as appears justified by each individual topic. I have been very fortunate in having access to the Smithers Rapra Polymer Library – a very comprehensive library which has a number of reports which I produced some years ago. Although some topics may read like a technical review, I have selected only sufficient information to make a point and not exhaustively included every reference. I am very grateful to many people for assisting me with this book: Frances Gardiner (iSmithers) for commissioning and co-ordinating the production. Steve Barnfield (iSmithers) for help, advice and preparation of figures and for typesetting the book and designing the cover. Elaine Cooper (iSmithers) for all her assistance in tracking down old reports. 1

Mixing in Single Screw Extrusion Sheila Cheese, Vicki Tweddle and Eleanor Carter (iSmithers) for sourcing journals and conference papers and John Colbert and Colin Chilles (Smithers Rapra) for loan of books. Ivan James for his advice on optical microscopy. Malcolm Davies who was involved in almost all the Rapra laboratory work described and who helped rescue a number of photographs. Hadj Benkreira of the University of Bradford, for loan of Richard Shales’ thesis. Lydia Cooper for turning my handwritten manuscript into a word document. Ken Gerber for information on the Metal Box mixer.

Martin Gale April 2009

2

1

The Need for Good Mixing in Single Screw Extrusion

In 1867 Tresca gave a paper at the meeting of the Institute of Mechanical Engineers titled ‘The Flow of Solids’ [1]. To a 21st century plastics engineer it comes as a surprise to find a 19th century publication that illustrates viscous laminar flow so clearly (Figure 1.1). The cross section in his figures 10 to 15 are remarkably similar to those in recent papers concerned with mixing in plastics extruders. However, Tresca did not make any suggestions for elimination of these laminar effects as he had deliberately produced them to demonstrate the behaviour of metals during rolling, forging, punching and planing. His technique was to ply discs of lead in a 100 mm diameter iron cylinder, insert a piston and force the lead through an iron die with a hole 50 mm diameter using an hydraulic press (a ‘giant’ capillary rheometer). His results were as follows: 1) The discs remain parallel in the cylinder outside the area affected by the ‘jet’ formation. 2) All the layers curve, converge, and bend over to form a ‘jet’. 3) The ‘jet’ is entirely composed of a cylindrical envelope formed out of the bottom disc of the original mass. 4) Each layer forms a distinct concentric tube. 5) Each layer is closed at the outer extremity by a more or less convex cap which is the central part of each of the original discs. The same results were obtained with tin, silver, aluminium and so on. Had he used thermoplastic discs, his results would have been much the same. In the extrusion of plastics products such as film, sheet, insulated wire, pipes and blow moulded containers and so on, the function of the extruder is to reliably produce a final product which meets the required specification at an economic price. 3

Mixing in Single Screw Extrusion

Figure 1.1 Diagram of flow of lead through dies from a paper presented by Tresca in 1865. (Reproduced from M. Tresca, Meeting of the Institute of Mechanical Engineers, Paris, France, 1867)

4

The Need for Good Mixing in Single Screw Extrusion To meet these requirements, final products often need to contain additives such as colour, antioxidants, slip, antiblock, flame retardants, tackifiers, fillers and so on, which have to be efficiently mixed into the polymer for the product to perform satisfactorily in service. Before examining the technology involved in mixing during single screw extrusion of plastics products, the extruder’s role in the overall scheme from individual materials to finished extruded composition needs considering. The majority of additives are powders and need to be in the form of very fine particles in order to perform satisfactorily and as a result tend naturally to agglomerate. A consequence of this is that mixing them into melted polymer requires finite forces sufficient to separate the individual particles and wet them with liquid polymer. Having achieved this, the particles must be further mixed to achieve a uniform concentration throughout the polymer. These two steps, which are described as ‘Dispersive Mixing ‘ and ‘Distributive Mixing’ are covered in more detail in Chapter 2, but are introduced here to fit the single screw extruder into the overall mixing picture. In general, single screw extruders are unsuitable for dispersive mixing and not efficient for distributive mixing either, but although little can be done regarding the former (with possible specialised exceptions in Chapter 13), very good distributive mixing is achievable. Mixing to achieve good dispersion is normally achieved using co-rotating twin screw extruders, continuous internal mixers, and batch mixers such as the Banbury mixer. Dispersive mixing in these machines is normally accompanied by good distributive mixing and hence these machines will produce well mixed plastics compounds ready for injection moulding and extrusion. This is shown in Figure 1.2(a). The demands on the product extruder are limited to pumping a fully melted polymer at a uniform temperature and economic rate through the die. The level of distributive mixing required to ensure melt homogeneity and temperature uniformity is normally achievable on modern extruders. The overall route may bypass the masterbatch stage depending on the type and level of additive. In the past, the routes in Figure 1.2(a) were very widely used, with many compounds available containing the right additives to meet particular needs and specifications. Over the years there has been a trend for additives to be incorporated at the product extrusion stage as masterbatches/concentrates in which the additives have been well mixed at a high concentration into a suitable polymeric carrier. These masterbatches, which are produced in similar equipment to that used for compounds, are added to natural polymer at a dilution which gives the required additive concentration in the 5

Mixing in Single Screw Extrusion

Figure 1.2(a) Operation for manufacture of extrusions – Compounding route.

final extruded product. (Figure 1.2(b)). The final extruder has an increased distributive mixing requirement of turning this pellet blend containing typically 5% (or even as low as 1%) masterbatch pellets into a uniform composition. Although this may need attention to mixing problems, it has a number of cost saving advantages for the product extruder. 1) Economies from buying masterbatch instead of compound, including reduced transport costs. 2) Reduced inventory costs. 6

The Need for Good Mixing in Single Screw Extrusion

Figure 1.2(b) Operation for manufacture of extrusions – Masterbatch route.

3) Flexibility to change natural polymer grade or supplier. 4) Responsibility for colour matching, technical support such as fire resistance testing etc. can be shared with the masterbatch supplier who has the necessary equipment, test facilities and expertise to supply customers on an individual basis. This enables the extrusion personnel to concentrate on running their plant as efficiently as possible. It should also be noted that rationalisation of the polymer industry has produced a continuing trend towards fewer and larger suppliers on a global scale who supply polymers as a commodity. As a result there is a shift in technical responsibility 7

Mixing in Single Screw Extrusion to plastics product extruders and injection moulders to produce extrusions and mouldings that contain additives necessary for the customers’ requirements. This also includes engineering plastics such as acrylonitrile-butadiene-styrene, a polymer widely used in coloured form for plastics sheet extrusion destined for applications such as thermo-formings. With extrusion companies so often caught in the middle between polymer suppliers increasing prices and customers demanding lower prices, economies in the efficient use of materials is of increasing importance. This may be achieved by: 1) Better mixing, enabling less additive to be used. 2) Production efficiencies from improved polymer blending and scrap recycling. 3) Retention of good mixing at increased output rate. Although the use of masterbatches confines the dispersive mixing role to compounding machines such as twin screw extruders, this route may demand a distributive mixing performance unachievable by many single screw extruders. This will be a particular problem when technical standards are required such as those for cables, or higher output rates are needed to remain economic. The single screw extruder not only lacks the flexibility of the screw configurations of the twin screw extruder, but the production situation is that product extrusion is limited to what exists at the time in the plant. The extruders may be ideal for the purpose. On the other hand, they may range in age and overall design and following past changes in product range have screws intended primarily for other applications. Fortunately, with an understanding of the various factors influencing distributive mixing, many screw design features and ‘add-on’ parts can be used to achieve the required results. Single screw extruders have been likened to ‘Grandma’s broom’ with replacement of screws, barrels, motors, bearings, instrumentation etc. as they become worn, unreliable or unserviceable. This can provide an opportunity for fitting new screws with improved melting and mixing performance, or adding melting/mixing devices to refurbished screws. One should be aware that replacing a worn screw which gives good mixing at low output rate as a result of poor pumping efficiency with a new one can result in high output rates with mixing inadequate for its present application. In their dispersive mixing role, co-rotating twin screw extruders have changeable screw configurations, multiple kneading sections, vacuum vents and downstream side feeders for fillers and fibres. Hence they are very adaptable to the mixing requirements of plastics compounding. However, these machines are complex and costly. They also tend to be limited to pellet production unless a gear pump or in-line 8

The Need for Good Mixing in Single Screw Extrusion single screw extruder is added downstream to generate the required die pressure to make, for example, sheet containing fillers or fibres. In comparison, single screw extruders are simple, rugged, low cost, low maintenance machines capable of developing whatever die pressures are required to make a very diverse range of extrusions. Consequently, there are ongoing and very diverse approaches into introducing ways of reducing the inherent dispersive mixing limitations of single screw extruders such that single pass extrusion might be used.

1.2 Examples of Mixing Problems The following examples illustrate a few of the wide ranging applications requiring better mixing by the finished product extruder.

1.2.1 Polyethylene Pipes and Cables In the 1970s, a number of polymer suppliers introduced a marketing policy which discontinued the supply of polyolefin compounds in favour of supplying natural materials. The discontinued compounds included black polyethylene for water pipe extrusion in which the pipe producers substituted the black compound fed to their extruders with pellet blends of natural polyethylene and carbon black masterbatch. Pipes made to technical standards applicable at the time [2-4] contained 2.5% carbon black suitably dispersed and distributed to provide protection for the polyethylene against the UV component of sunlight. The advantage offered to pipe producers for using a blend of natural polyethylene and carbon black masterbatch was a very significant materials cost savings. Unfortunately the extruders at that time experienced difficulties in meeting the mixing requirements of the applicable standards. Those standards have since been incorporated into current European ‘harmonised’ versions [2-4] in which the carbon black dispersion requirements are essentially the same. Typical cross sections are shown in Figures 1.3 and 1.4 (Figure 1.4 is a negative of Figure 1.3 and shows the striations more clearly). Photomicrographs have also been shown in connection with meeting the outdoor performance application standards for insulated cables by Patch [5], whilst Lee and Borke [6] have quite recently included mixing effects among factors affecting the performance of carbon black masterbatches in wire and cable applications. 9

Mixing in Single Screw Extrusion

Figure 1.3 Pipe cross section photomicrograph without breaker plate. (Photograph taken by D.I. James. ©Rapra Technology)

Figure 1.4 Pipe cross section photomicrograph (negative) with breaker plate. (Photograph taken by D.I. James. ©Rapra Technology) 10

The Need for Good Mixing in Single Screw Extrusion

1.2.2 Blow Moulded Bottles Figure 1.5 shows a blow moulded bottle which formerly contained a laundry softener. Viewed through the bottle’s neck there are stripes of poorly distributed blue masterbatch. This indicates that the blow moulder had ‘given away’ expensive colour masterbatch. If good distributive mixing had been achieved by the extruder, less masterbatch could have been used and the bottle more profitable. Masterbatch mixing appears to be a particular problem with blow moulded containers in high molecular weight polyethylene [7, 8].

(a)

(b)

Figure 1.5 Blow moulded bottle. (a) complete bottle, (b) Inside view through neck showing masterbatch stripes. 11

Mixing in Single Screw Extrusion

1.2.3 Chalk Filled Polypropylene Pipe Rigid pipe was required, extruded as a blend of polypropylene pellets and 40 wt% surface treated Calcium carbonate filler as a substitute for rigid polyvinylchloride (PVC). The speckled pipe sample (Figure 1.6) clearly shows that dispersive mixing of fine powders is difficult or impossible with a single screw extruder under normal conditions. A pipe with the same composition was produced using a very high concentration of chalk masterbatch prepared in a twin screw extruder and let down to 40 wt% chalk in the single screw extruder used previously. In this case the dispersion (and distribution) was good, but the two stage process would have been uneconomic.

(a)

(b)

Figure 1.6 Polypropylene pipe with (a) undispersed agglomerated filler; (b) produced using a two-stage process 12

The Need for Good Mixing in Single Screw Extrusion

1.2.4 Blown Film Film blowing will show up problems of poor masterbatch distribution even more than blow mouldings and will also show the presence of agglomerates due to substandard pigment dispersion. This technique is sometimes used on a laboratory scale for pigment masterbatch quality control. The addition of a mixing element of the types described in Chapter 9 will often enable the extruder to produce good striation free film from the same polymer and colour masterbatch blend which otherwise produces film with coloured stripes.

1.2.5 Industrial Blow Mouldings Blow moulded containers forming part of a machine incorporated a flame retardant masterbatch in order to meet a UL94 spread of flame requirement which used test pieces 12.7 mm wide by 150 mm long. The results for individual test pieces ranged from immediate flame extinction to burning the full length of the test piece: most results being scattered between these two extremes. Microscopy examination revealed that the scatter of results was caused by the flame retarding masterbatch being distributed in bands similar to those for bottles described in Section 1.22. It appears that the blowing process exaggerates this effect. Inadequate mixing of flame retardant masterbatches has also resulted in corner cracking of large blow moulded drums used for transporting chemicals when drop tested. Following laboratory trials to evaluate potential mixing devices a mixer was retrofitted as a screw extension which solved the problem.

1.2.6 Production Scrap Re-use Distributive mixing can influence production economics in situations other than the incorporation of masterbatches and additives. In the production of thermoformed food packaging containers, thin sheet edge trim and skeletal scrap (a continuous sheet full of holes following separation of formed containers such as round yoghurt pots) can represent 40% of the original sheet. It is essential to the economics of the process that this scrap is recycled. In an example, a five layer barrier sheet consisting of a barrier polymer layer with adjoining adhesive and polyolefin layers on either side had an additional layer of ‘buried scrap’. Inadequate homogenising of the separate components of the scrap layer resulted in unacceptable ripples in the thermoformed pots. Following laboratory extrusion trials with the granulated scrap, a mixing device was fitted to the scrap layer production extruder which solved the problem. 13

Mixing in Single Screw Extrusion

1.2.7 Agglomerates and Gels in Thin Extrusions The presence of undispersed particle clusters in thin sheet and film can result in holes or thin lines prone to splitting in use. Normally an agglomerate will be found at the edge of the hole and at, or near, the start of the thin section line. These particles may be additive agglomerates, gels, or contaminants. The agglomerates should not be present in the feed material. Gels may result from a need to locate ‘hang-up’ areas in the machinery to prevent oxidation or if a melting problem it might be dealt with by dispersive mixing. In practice, this is difficult (see Section 14.11).

1.2.7.1 Thin Plasticised PVC A very thin flexible PVC extrusion proved unsatisfactory in service due to porosity. A microscopic examination showed that the holes had been generated by pigment agglomerates, each hole having an agglomerate at its edge. Initially the extruder had been blamed for a lack of mixing, but the remedy was entirely with the compound supplier to ensure the PVC compound was free from agglomerates in the first place. It was important in this case to establish that the problem was caused by pigment, filler, or stabiliser and not degraded PVC or contaminants such as dirt off the floor following repacking a split bag by a carrier. The use of fine wire mesh screens at the barrel exit by the product extruder may suffice to catch a very low concentration of agglomerates, paper wrapping, and ball point pen tips, but might not be used with PVC in case of stagnation causing thermal decomposition.

1.2.7.2 Silage Wrap Splitting in Use Following the rotting of silage, it was found that considerable amounts of film had split, allowing rain to penetrate the silage during outdoor storage. Examination showed particles existing at the start of each slit, which appeared to be carbon black agglomerates. However, infra-red analysis showed the particles to be crosslinked gels, most likely caused by thermal oxidation during the extrusion process.

1.2.7.3 Holes in Silage Wrap Film During Film Blowing Holes caused by gels appeared during film blowing of linear low-density polyethylene silage wrap film. The gels appeared to melt when film samples were heated on a microscope hot stage, but infra-red analysis showed the observed gels were clusters of 14

The Need for Good Mixing in Single Screw Extrusion finely divided partially crosslinked particles. Suspecting that the problem was caused by oxidative crosslinking in the extrusion equipment, an antioxidant masterbatch was blended with the natural polymer pellets and a liquid antioxidant was added to the polybutene tackifier being pumped into the extruder. Although this increased materials cost, it transformed the film quality. In addition to eliminating formation of large gels previously causing the holes, the large number of smaller gels was also eliminated. In these two silage wrap film examples, the problems lay with the formation of gels which ideally would be destroyed by good dispersive mixing. However, when such gels consist of partially crosslinked rubbery unmelted oxidised particles, the solution widely used in blown film extrusion is to filter out the particles using fine mesh screens, although smaller gels may deform like soft rubber balls, squeezing through the screen apertures and recovering their shape after the screen. Unfortunately the gels may develop after the screens both in the die or more likely as a result of co-extrusion feed pipes. These may be long with ‘dog legs’, corners, and no chrome or polished surfaces in difficult to access areas. The answer to the problem of gels is to avoid their formation in the first place if at all possible (see Chapter 14).

1.2.8 Transparent Polycaprolactone/SAN Blends Polycaprolactone, which is widely used in medical applications, can be blended with a number of polymers such as styrene-acrylonitrile (SAN), PVC, and polycarbonate. In this example a polymer blend of polycaprolactone with a high nitrile SAN was expected to give a transparent extruded sheet which was thermoformable in hot water. Suitable thermoforming properties and adequate transparency had been achieved with 35 wt% polycaprolactone blended with 65 wt% SAN using small laboratory samples prepared in a torque rheometer. Unfortunately, strips extruded from a pellet blend using a 25 mm laboratory extruder were white, cloudy and not transparent. Following work with an independently driven mixer, subsequent trials with a 38 mm extruder having a cavity transfer mixer attached as a screw and barrel extension gave acceptable transparency.

1.2.9 Decorative Wood Grain Effects A lack of distributive mixing can be exploited, providing it is controlled, to give decorative patterns on the surface. This is normally achieved using blends of different coloured pellets which also differ in viscosity. The patterns are the result of 15

Mixing in Single Screw Extrusion a combination of the melting process producing coloured ribbons and viscous flow in the die with drag at die surfaces producing coloured striations as described in later chapters. This effect can be exploited to provide extruded wood grain effect plastic profiles for office furniture, shop fittings, etc. Special masterbatches are available for such applications. This continues a long history in which coloured blends of cellulosic plastics were developed for spectacle frames, pen barrels and combs: processes which originally exploited the poor mixing performances of the short barreled extruders and plunger injection moulding machines of the past.

References 1.

M. Tresca in Iron and Steel Manufacture, Ed., F. Kohn, William Mackenzie, London, UK, 1868.

2.

BS EN 13244-1, Plastics Piping Systems for Buried and Above-Ground Pressure Syetems for Water for General Purposes, Drainage and Sewage – Polyethylene (PE) – Part 1: General, 2003.

3.

BS EN 13244-2, Plastics Piping Systems for Buried and Above-Ground Pressure Systems for Water for General Purposes, Drainage and Sewage – Polyethylene (PE) – Part 2: Pipes, 2003.

4.

BS EN 13244-5, Plastics Piping Systems for Buried and Above-Ground Pressure Systems for Water for General Purposes, Drainage and Sewage – Polyethylene (PE) – Part 5: Fitness for Purpose of the System, 2003.

5.

R. Patch, Kunststoffe, 1975, 65, 2, 89.

6.

C.D. Lee and J.S. Borke in Proceedings of the 62nd Annual SPE Conference – ANTEC 2004, Chicago, IL, USA, 2004, p.288.

7.

D. Boes, Kunststoffe, 1974, 64, 11, 641.

8.

G. Martin, Kunststofftechnik, 1972, 11, 12, 329.

16

2

Dispersive and Distributive Mixing

2.1 Definitions and Illustrations These terms were introduced in Chapter 1, but in considering the opportunities and limitations of exploiting the single screw extruder as a mixer during product extrusion, discrimination between these two types of mixing is essential. Dispersive Mixing: An operation that reduces an agglomerate size of the minor constituent to it’s ultimate particle size. Distributive Mixing: An operation that increases the randomness of the spatial distribution of the minor constituent within the major base with no changes in the size of the minor particle. The two operations are shown in Figure 2.1. Hopefully, no readers of this book are confronted with this situation! It does, however, illustrate that large forces may be necessary to achieve good dispersion (Note: large size of motor, in addition to the pick axe!). Also for good distributive mixing, high shear strain (a lot of whisking!) is required. A simple illustration of these two terms is shown in Figure 2.2. Each of the four diagrams has the same number of spots to represent additive particles at their ultimate size. Figure 2.2(a) shows bad dispersion and bad distribution. There are many agglomerates and also areas of high and low concentration. Figure 2.2(b) shows bad dispersion and good distribution. The agglomerates are uniformly spaced. Normally this would be an undesirable situation, but (within limits), might be desirable for carbon black in an electrically conducting extrusion. Increasing dispersive mixing will reduce conductivity. Figure 2.2(c) shows good dispersion. All the agglomerates have been broken down into ultimate particles which are separated and surrounded by polymer. However, distribution is poor with areas of high, low and zero concentrations. In a real single screw extruder situation this product would most likely have laminar striations. Figure 2.2(d) shows good dispersion and good distribution. 17

Mixing in Single Screw Extrusion

Figure 2.1 Distributive mixing and dispersive mixing. (Reproduced with permission from Richard Juniper. ©2009, Richard Juniper)

Figure 2.2 Diagrammatic representation of dispersive and distributive mixing effects. 18

Dispersive and Distributive Mixing

Figure 2.3 Photomicrographs showing real mixing situations when incorporating carbon black masterbatch to give 2.5% carbon black in polypropylene. (©Rapra Technology)

Real situations for the same material composition are shown in the three photomicrographs of Figures 2.3(a), (b) and (c). All three are for the same pellet blend of polypropylene and carbon black masterbatch giving a final carbon black concentration of 2.5%. This is typical for an extruder feed for a product to be used outdoors and consequently requiring protection against the UV component of sunlight. In Figure 2.3(a), a single screw extruder with a general purpose screw with no mixing devices produced numerous laminar streaks and there are carbon black agglomerates present, although not easily identified amongst the striations of masterbatch. In Figure 2.3(b) a cavity transfer mixer (CTM) extension (see Chapter 9) has been added to the screw used in Figure 2.3(a) which has eliminated the striations but agglomerates are still present and in the absence of striations are more clearly seen. The greyness gradation is the result of microtomed thickness variation. In Figure 2.3(c), a barrier melting/mixing screw was used and shows a mixture of striations and agglomerates. (Note that carbon black agglomerates are allowed by pipe and cable standards providing their number and size are within specified limits). As described in Chapter 7, the prime function of barrier screws is to control melting. Mixing will often be improved as a result of this function since it is dependant on melting as explained in later chapters. 19

Mixing in Single Screw Extrusion There are a confusing number of terms used to describe mixing. ‘Distributive mixing’ may be described as ‘simple mixing,’ ‘extensive mixing,’ or ‘blending’. whilst ‘dispersive mixing’ may be described as ‘intensive mixing’ and ‘elongational mixing’. Furthermore, dispersive mixing is usually accompanied by distributive mixing but the opposite does not apply. Many of the terms describe the mechanism involved, such as elongation, (or stretching), squeezing (or kneading), laminar shearing, etc. As this book is concerned with single screw extruders, most of the mixing which takes place is ‘distributive’ and accomplished by laminar shearing. In comparatively recent developments, dispersive mixing in single screw extrusion has been investigated for polymer blending (Chapter 14). As droplet mixing has traditionally been considered a dispersive mixing process, methods of mixing incompatible polymers using (dispersive) elongational mixing are usually adopted. However, it has been found that although the mixing mechanisms which occur in polymer blending can be surprisingly complex, they are achievable using single screw extruders (see Chapter 14).

2.2 Dispersive Mixing 2.2.1 Dispersive Mixing Mechanisms Dispersive mixing is required for the mixing of solid additives into polymers. Most additives are solids and for effective performance need to be used as fine powders and well mixed into the polymer. Additives with a particle size of less than 100 µm tend to naturally agglomerate. These agglomerates will often be present, no matter what precautions are taken, whilst the mixing processes used can sometimes contribute to the problem. The attraction forces between particles require tensile forces sufficiently high to separate them and cause break up of the agglomerates and also promote wetting by the polymer of the newly exposed particle surfaces. To achieve this, the surrounding polymer or polymer/wax matrix needs a flow field which will provide the necessary force. The turning rotors of internal mixers and co-rotating twin screw compounding extruders will usually provide the required shear forces. The overall mixing mechanisms of these machines is very complex [1, 2], but the fundamental behaviour can be represented by the following simple model. In Figure 2.4.1(a) an agglomerate X enters the nip formed by the tip of the turning rotor and mixer wall. There will be a combination of compression, shearing and acceleration subjected to the agglomerate X as it passes into and through the gap 20

Dispersive and Distributive Mixing

Figure 2.4 Dispersive mixing in an internal (batch) mixer.

formed between the rotor tip and chamber wall. This will be followed by stretching forces from the diverging flow paths in the gap exit region which may also rupture particles (Figure 2.4.1(b)). These then move around the mixer following complex paths dictated by the machine design. They will pass through a nip (maybe a different 21

Mixing in Single Screw Extrusion

Figure 2.5 Influence of four mixer speeds at three mixing times. (© R. Butterfield, University of Bradford, UK) 22

Dispersive and Distributive Mixing one) separately for a second time, breaking up further: the process being repeated until break down to the ultimate achievable particle size is reached (Figures 2.4.2(c)2.4.7). It is likely that the orientation of the clusters will change each time they enter the high stress field, which increases the chances of rupture. The complex paths taken by the particles also achieves good distributive mixing The degree of dispersion, i.e., reduction in particle size achieved, will depend on rotor speed and mixing time. Increasing the speed increases the stresses applied to the agglomerates and the frequency of passes. In Figure 2.5, bar charts show the particle size distribution for 4 rotor speeds in a ‘Minimixer’ [3] for a colour masterbatch. Three runs of 4, 8 and 12 minutes were made at each of the 4 speeds. (Due to limited space the data is confined to the 0.60 to 0.010 particle size range). These results clearly illustrate how the number of small particles increases with increasing speed and time. It is also interesting that the peak is skewed so close to what may be the ultimate particle size. Dispersion forces will also be increased by increasing viscosity. This may be achieved by using minimum temperatures, higher viscosity polymer or higher pigment or other additive levels, but this is complicated by requirements of particle wetting and air displacement. Some pigments may be damaged (‘bruised’), resulting in a change of colour. The effect of the elongational flow field for dispersive mixing (although the geometry was quite different), was demonstrated in model experiments by Theodorou [4]. The apparatus was based on that used originally by Taylor [5] for studying droplet behaviour and compared simple shear flow with hyperbolic elongational flow using model agglomerates instead of droplets. The apparatus had four driven shafts arranged in a square which had alternative arrangements to demonstrate both the elongational flow field used to achieve dispersive mixing and laminar shear associated with single screw extruder mixing. The device was immersed in a viscous silicone fluid in a small glass aquarium. A spherical agglomerate consisting of acrylic powder bound together with silicone fluid and having a solids concentration of about 70% was placed at the centre of 4 rollers located on the spindles. The rollers were then rotated to apply an elongational flow field to the agglomerate as shown in Figure 2.6. A particle cloud was formed which became aligned in the direction of maximum stress whilst the agglomerate was subjected continuously to the maximum elongational force. To compare elongational flow with shear flow, the rollers were replaced with sprockets and two endless belts made from 35 mm film were fitted. By rotating all the sprockets in the same direction, a laminar shear flow field was generated. By having two surfaces moving in the opposite direction, the agglomerate remained in the field of view. Under these conditions the agglomerate rotated, subjecting the particle interfaces to alternating tension and compression. 23

Mixing in Single Screw Extrusion

Figure 2.6 Model experiments illustrating laminar and elongational shear fields.

The elongational flow field was found to be two to three times more effective for dispersing than the shear flow field. A figure of 6 times has been given for experiments using cohesionless particle clusters [6, 7].

2.2.2 Dispersive Mixing of Additive Powders Such as Pigments Additives required to give properties for specific applications such as pigments, flame retardants, fillers etc are almost always in the form of very fine powders as this is usually necessary for them to perform effectively. Although this can cause problems (with the single screw extruder normally being unable to satisfactorily incorporate these additives), they are more usually the concern of the compound and masterbatch supplier. However, it is useful for the product extruder to appreciate the factors involved in dispersive mixing of powders into polymers. There are many situations that may result in the presence of agglomerates which cause trouble whether from 24

Dispersive and Distributive Mixing appearance or performance, or both. An understanding of the factors involved gives an appreciation of why good dispersive mixing of powders is difficult to achieve in single screw extruders, and also helps to remedy problems where suppliers are involved. There is also the ongoing desire to use single screw extruders in a compounding role, ideally combined with product extrusion such as filled sheet or pipe; for example, to compete with rigid polyvinylchloride (PVC). With a growing need for recycling post consumer waste, the single screw machine’s simplicity and ruggedness is more suited to meeting uncertainties of feed stock compositions and potential contaminants. It may also make an otherwise uneconomic situation viable by combining compounding with product extrusion. Pigments are probably the most widely used additive and generally the most demanding. The customer may well be critical of the smallest discrepancy in colour shade, opacity or transparency, gloss etc. particularly where matching with other components is necessary such as painted automotive parts and thermoformed bathroom items and so on. The dispersion process as a whole for the incorporation of fine particle pigments and fillers etc into polymers contains a number of obstacles inherent in the physical requirements of the process. The ideal requirements for polymer and additives, together with the reality are shown in Table 2.1.

Table 2.1 Comparison of ideal and real conditions for good dispersion Property

Requirement for good dispersion

The reality

Particulate size

Similar

Very different

Surface energy

Similar

Very different

Very low

Very high

Polymer viscosity

1. The finer the particle size of the pigment, flame retardant etc. then normally the better its final performance, but with particle sizes of less than 100 µm agglomeration naturally occurs. To minimise agglomeration, polymer powder should be used, [8-12] but polymers are normally supplied as 3 mm pellets. The range of reactor powder (except for PVC) is very limited and grinding increases costs. 25

Mixing in Single Screw Extrusion 2. Pigments, fillers, and flame retardants are often polar and hydrophilic (even organic pigments can contain water of crystallisation), whereas polymers tend to be non-polar, hydrophobic hydrocarbons. 3. Polymers normally have high melt viscosities (in contrast to oils used for paint pigment dispersions) which makes penetration between additive particles difficult. For pigments, the opacity, transparency and gloss can be related to particle size [9]. For example, efficient hiding power with titanium dioxide pigment is achieved with a particle size of about 0.25 µm. For particles of about 100 µm and below, the attractive force exceeds the separation forces, which is why powders are free flowing above this diameter and cohesive if less. The fine grinding used to produce very fine powders promotes a tendency to spontaneously agglomerate to reduce surface energy. In a very comprehensive set of articles on the importance of the physical properties of pigments in relation to their dispersion into plastics, Smith [8-11] described a range of circumstances which could cause agglomerates to be formed. In addition to promoting a tendency for spontaneous agglomeration by fine grinding, further agglomerating factors were impurities, dust preventing additives and compaction of these very low bulk density materials. These were either deliberate for transportation or unintentional from silo storage or stacking bags. During mixing of pigments with polymer pellets, weak agglomerates may be dispersed, but agglomerates may be formed by compaction between plastics pellets and between pellets and the mixer wall. Increasing the polymer surface area by reducing the pellet size will reduce the pigments tendency to form agglomerates but may leave weak agglomerates undispersed. High speed mixing can substitute one agglomeration mechanism for another. Impingement with the rotor blades may disperse particles, may cause agglomeration or may result in pigment build-up on the mixer blades of up to 20% of the pigment. This may later break off during mixing and produce further agglomerates. Smith’s recommendation was that the best way to avoid agglomerates was to prevent their formation in the first place [8]. In dispersing powders, they must be wetted by the matrix to displace air from polymer clusters such as agglomerates and coat each particle. This is more easily achieved by the comparatively low viscosity of paints than the high viscosities of molten plastic. However, the former requires higher standards of dispersion. Further requirements are the removal of water, air and other impurities at the particle-polymer interface and the subsequent establishment of a firm bond between the two phases [12]. Smaller particles also contribute to a higher colour value. Some pigments have hydrophobic surfaces whilst others are hydrophilic. Consequently the former are more easily dispersed in organic media of low polarity such as hydrocarbon plastics, but the latter will be wetted more easily by water and highly polar media. For example, Phthalo Blue, which is hydrophobic, is used in plastics, 26

Dispersive and Distributive Mixing elastomers, paints and inks. But Ultramarine which is hydrophilic is used in paints and inks but not plastics and elastomers [11]. Surface active agents help dispersion and prevent agglomerate formation by reducing interfacial surface tension, lubricating surfaces to increase mobility and prevent agglomeration. Adsorbed water and other impurities will be removed which otherwise would hinder dispersion. The following economic losses caused by pigment dispersion problems were described by Smith as follows [8]: 1) Extra production time needed to achieve satisfactory dispersion. 2) Expensive pigment wasted when extra pigment is added to compensate for the disproportionate quantity of total pigment which comprises the agglomerates. 3) Technical problems caused by presence of agglomerates, e.g., electrical faults, reduced mechanical strength, breakage of fibres and films during production, anomalous weathering resistance and colour variations. 4) ‘Off-shade’ product when one pigment of several in a blend forms agglomerates, plus colour variations during production runs. 5) Rejection of finished products having obvious pigment defects and streaks. Smith compressed a range of pigments in a tableting machine instrumented to measure radial and axial pressures and concluded that the more difficult to disperse organic pigments, (particularly phthalocyanines), were more elastic than inorganic ones. Some (mainly organics), were more prone than others to stick to the metal mould, suggesting they were more likely to stick to mixer impellers [8]. Crushing tests showed large differences in compaction strengths between organic and inorganic pigments. Pigment dispersability is also influenced by its relative hardness/softness, manufacturing route and surface treatment [13]. Fillers and flame retardants such as alumina trihydrate are frequently used at concentrations of up to 50 wt%. Being of a similar concentration to a pigment or carbon black masterbatch, they are prepared like masterbatches in internal and twin screw compounding machines, but often with no dilution required in the final product extrusion. All fillers increase stiffness, decrease elongation and thermal expansion and improve dimensional stability of extrusions and mouldings. However, whether they act as a ‘filler’ or ‘reinforcement’ depends on the aspect ratio of the particles. A simple 27

Mixing in Single Screw Extrusion substitution of a significant proportion of the ‘expensive polymer’ with a ‘cheap filler’ to reduce final overall extrudate cost is not readily achievable following density considerations and compounding costs. As a result, there is an interest in single screw one-pass extrusion, similar to unplasticised PVC powder dry blend extrusion of building products with twin screw extruders, but using the cheaper, simpler single screw machine (see Chapter 13). Schlumpf [14], has listed eight factors to be considered by the processor, of which four were most important: 1) Particle shape. 2) Particle size distribution curve with ‘top cut’. 3) Specific surface. 4) Surface energy. Although similar considerations apply to pigments, with the combination of high concentrations and reinforcing role, certain particle characteristics may be of greater concern than with many other additives. Smaller filler particles which have the potential to give the best reinforcement due to higher interaction with the polymer may provide dispersion problems and unprocessable high viscosity compositions. Too high a surface energy may cause dispersion problems which negate the potential advantage of high interaction of filler and polymer, although filler manufacturers can apply coatings which reduce the surface energy and thereby improve dispersion. Surface energy figures for a range of fillers have been listed by Schlumpf [14]. Approximate ratios for some of these values to that of typical plastics are shown in Table 2.2.

Table 2.2 Approximate ratios of surface energy values for several fillers/ reinforcements and titanium dioxide compared with an average value for plastics Filler/Reinforcement

Ratio to typical plastics

Mica

65-145

Glass

30

Titanium dioxide

15

Calcium carbonate (chalk)

2

Talc

2

Plastics

1

28

Dispersive and Distributive Mixing

2.3 Distributive Mixing A uniform spatial distribution of additive particles is normally achieved concurrently with good dispersive mixing in internal mixers and twin screw compounding extruders. Consequently masterbatches and compounds contain additive particles which are uniformly spaced and free from laminar streaks. No further particle mixing is necessary for compounds, but with masterbatches, the single screw extruder will be required to achieve a dilution to a completely uniform composition during extrusion, i.e., provide good distributive mixing. Although dispersive mixing is rarely possible in single screw extruders, distributive mixing adequate for the application should be achievable, providing the extruder is suitably equipped to meet the level of mixing required for the product. In the single screw extruder, the viscous polymer melt undergoes a mechanism of laminar flow in which the shear stresses are determined (and therefore limited) by the viscosity of the polymer, but the total shear strain available is restricted by the length of the extruder. However, although there are limitations to this generalised assumption, which may result in an unsatisfactory result, there are numerous approaches and devices available which, to varying degrees (and cost), can succeed in overcoming them. There are also variables within the extrusion process which often need to be taken into account. Every stage in the extruder from the pellet feed to the die, should be considered as all have an influence on the level of mixing which will be achieved. On the other hand, with an understanding of single screw distributive mixing, novel processes are possible which can provide significant technical and economic benefits.

2.3.1 Laminar Shear Flow Mixing Laminar shear occurs in viscous fluids situated between fixed and moving surfaces, e.g., between an extruder barrel surface and rotating screw [15] (Figure 2.7). The behaviour of such liquids will be very much influenced by viscosity. Viscosity is defined as the ratio of shear stress to rate of strain (Figure 2.8). As molten polymers are high viscosity fluids, they cannot be readily subjected to turbulent flow, which otherwise would be a very efficient method of mixing from a practical standpoint. Neither will they readily mix by molecular diffusion like gases and low viscosity liquids. Being viscous, molten polymers continue to deform as long as a stress is applied. For 29

Mixing in Single Screw Extrusion

Figure 2.7 Laminar shear flow between stationary and moving plates.

Figure 2.8 Viscous behaviour of molten polymers. 30

Dispersive and Distributive Mixing

Figure 2.9 Melt flow index tester: Constant shear rate and laminar flow.

example, in a melt flow index tester the molten polymer will flow through the nozzle at a constant rate when the weight is applied to the piston (Figure 2.9). The energy generated by the deformation is dissipated as heat. This is an important processing characteristic as rapid stressing such as high extruder screw rotation speeds can, in some circumstances, generate heat faster than it can be removed, resulting in thermal degradation which produces discoloration, gels, inferior mechanical properties, etc. 31

Mixing in Single Screw Extrusion

Figure 2.10 Viscosity of molten HDPE at several temperatures. (Reproduced with permission from D. Fleming, Fleming Polymer Testing and Consultancy. ©D. Flemming)

If the molten polymer flows with a viscosity independent of stress level, its behaviour is said to be Newtonian. However, within the screw channel, the long entangled polymer molecular chains will become partly unravelled and become aligned in the direction of shear. An analogy used by Cogswell [15] was the generation of a more ordered state of spaghetti on a plate following twirling with a fork. This molecular alignment reduces viscosity and the shear thinning is termed pseudoplastic (Figure 2.9). To provide a viscosity ‘picture’ of a particular polymer grade, it is normal practice to plot viscosity against shear rate on a log/log scale and include graphs for a range of temperatures (Figure 2.10). The reduction in viscosity by shear thinning is important in screw extrusion as it constrains die pressures and partly mitigates the shear heating effects caused by increasing screw speed. If this were not the case, thermoplastics would be 10 to 1000 times more difficult to extrude.

2.3.2 Measurement of Distributive Mixing Achieved by Laminar Shearing A model used to demonstrate laminar shear mixing is ‘Couette flow’ [16, 17], the principle used in the ‘Couette Viscometer’ The geometrical arrangement is to have the two viscous fluids to be mixed occupying two halves of an annulus formed by a rotatable mandrel within a cylinder (Figure 2.11). 32

Dispersive and Distributive Mixing

X

Y

Figure 2.11 Couette flow.

As viscous polymers normally wet the surfaces of plastics machinery, we assume the two fluids X and Y will remain in contact with the surfaces of the respective halves on the outside and on the mandrel. As the mandrel is rotated, the area of contact between the two fluids increases in proportion to the number of rotations of the mandrel. This increase in area can be used as a measure of the degree of mixing:

Degree of mixing =

A new area = f original area A o

The greater the number of rotations, the greater the increase in area, and hence the ‘degree of mixing’. As the volume of fluid remains constant, the striation thickness must decrease at the same rate, and with the volumes remaining constant, the original thickness divided by the new thickness will give the same answer:

t A Final area original thickness = or f = o Original area final thickness Ao tf For measurement of distributive mixing, striation thickness is a far more practical proposition than measuring surface area (see Chapter 3). In the above example the 33

Mixing in Single Screw Extrusion annulus contains two fluids in equal amounts, but the same situation applies when one component forms a very small proportion of the total area, as would be the case for 5% of a 40% carbon black masterbatch. A simple demonstration is to add a small amount of fruit concentrate or jam to natural yoghurt and stir slowly, observing how the concentrate forms striations which become longer and thinner. It also demonstrates the surprising amount of stirring necessary to produce a uniform colour and the impossibility of dispersive mixing bits of fruit, skin and pips with the very limited forces available by stirring.

2.3.3 Limitations of Lamina Flow Mixing In considering the laminar shear flow due to screw rotation, we can simplify the mechanism by considering a cross section as follows (Figure 2.12).

Figure 2.12 Laminar shear in the screw channel. 34

Dispersive and Distributive Mixing Movement of the screw surface shears the viscous polymer melt in a similar manner to the Couette viscometer. Using the scheme described by McKelvey [16], a striation A B C1 D1 will be sheared as the screw rotates to position A B C2 D2 which can be defined by the angle ∅. Shear strain by definition is the amount moved divided by the distance between the two planes: Shear strain (ϒ) = L/h = Cot ∅

Striation thickness ratio =

t1 = sin ∅ t0

Where t1 = original striation thickness t2 = striation thickness after applying a shear strain of L/h L = length h = height From a starting position of zero strain i.e., L = 0 and ∅ = 90° to a position where strain is high such that ∅ is correspondingly very small, i.e., approaching zero, we can plot Sin ∅ against Cot ∅ to show how striation thickness ratio changes with shear strain (Figure 2.13).

Figure 2.13 Limitations of laminar shear mixing in a single screw extruder. 35

Mixing in Single Screw Extrusion The overall effect is that most of the mixing, as measured by striation thickness ratio, occurs during initial straining and then rapidly tails off so that further strain has little or no effect. The striation behaviour shown in Figure 2.12 between a moving and stationary surface is a very simplified picture of the real situation where striations spiral down the screw channel. However, the alignment of striations observed in screw channel samples demonstrates that the simple model is valid. Following melting, in which masterbatch striations are formed, the mixing efficiency rapidly declines. As a result the remaining length of the screw channel has very little effect on mixing these aligned striations. The overall position is that once these striations are aligned, the shear strain available in the remaining length of the screw channel before the die is reached, has very little effect. As a consequence, there may well be striations in the extruded product as described in several of the examples in Chapter 1.

2.3.4 Eliminating Laminar Striations As stated by Smith [18]: ‘There is little more to mixing in polymers than meets the eye’. The resolution of the human eye, is approximately of the scale 10-4 m, and therefore, the ultimate final mixing goal in many applications. It is possible that by the time the screw tip is reached, masterbatch striations, although subjected to very limited shear mixing following orientation in the direction of shear, will be adequately mixed for the application. However, there are a number of other factors which need to be considered including the following: 1) Lateral stretching during blow moulding and film extrusion emphasises the presence of striations. 2) Variable weathering performance and flame propagation. 3) Possible reduction in mechanical properties. 4) Increased materials costs from inefficient use of masterbatch.

2.3.4.1 Fundamentals Erwin [19] commented that it should be possible to take advantage of the predictable laminar orientation to design mixing sections that increase the rate of mixing. From observations of shear flow behaviour it appeared reasonable to assume that one can rely on a consistent orientation which is independent of the initial orientation. Mixing sections should then be designed to rotate the fluid and present the striations at the most favourable orientation for the following section to continue the shearing 36

Dispersive and Distributive Mixing process. Furthermore, such rotations of a viscous fluid involve minimal energy in comparison to shearing. He further reasoned that incorporating a series of mixing sections which rotated the striations into the most favourable position for the following shearing stage, would impart mixing efficiency. This would require the interfaces to be vertical to the shear plane after re-orientation, i.e., turned through 90°. Referring to the graph in Figure 2.13, the effect is to repeatedly return from a point somewhere beyond Y and return to X. Erwin also included a table showing a theoretical optimum number of re-orientated shearing stages necessary to produce the mixing equivalent to an extruder without re-orienting devices imparting a shear of 10,000 [20]. The optimum was 10 shearing sections (i.e., 10 - 1 = 9, turning elements) which reduced the overall shear from 10,000 to 42, but even two shearing sections reduced the necessary shear to 280. The effectiveness of this mechanism for achieving good distributive mixing was clearly demonstrated in experiments by Ng and Erwin [21] based upon the concepts described by Spencer and Wiley [22]. In the distributive mixing of two similar viscous lamina flow liquids, the degree of mixedness can be assessed by either the total interfacial area between them, or the striation thickness, the two being related:

Interfacial area × mean striation thickness = volume of liquid 2 Referring to Spencer and Wiley [22]: growth of interfacial area in a fluid subjected to shear follows the formula:

A = 1 – 2S Cos α Cos β + Cos2 α S2 Ao Where A is the new area

Ao is the original area, and



S is the magnitude of shear strain



α and β are angles defining the orientation of the surface to the shear strain [21]. 37

Mixing in Single Screw Extrusion For large unidirectional simple shear, as found in the metering section of a single screw extruder:

A = S Cos α Ao The growth of interfacial area proceeds linearly with shear. Erwin’s model [20] for a mixing section with flow interruption by pins etc. considers the extruder to subject the melt to large unidirectional shear and then distort it to randomise the orientation and then subject it to further shear. Since there is no energy dissipation in the solid body rotation of a viscous liquid, the energy expended in this re-arrangement is considered as being small in comparison to the work applied in the shearing section. As large shear orients interfaces parallel to shearing planes, further improvements in mixing efficiency would be attained by each mixing section rotating the fluid so that the interfaces are more favourably oriented for the next shearing stage.

A = S Cos α this will occur when Cos α = 1. Ao If a shearing system has many mixing stages dividing the system into sections having large shear with equal magnitude, mixing will be:

In the equation

N−1

Af  S  S  =   A o  2N   N 

N

S = 12   N

N

A S With an input of optimal orientation f =   Ao  N 

A very simplified approach is to consider a screw tip mixer with well-spaced rows of pins. We then assume that between the rows of pins, simple shear is applied and in passing through each set of the pins the melt is turned and repositioned at right angles to the direction of shear applied between this row and the next row of pins and so on (Figure 2.14). (Although rows of pins work moderately well, some form of interaction with static pins is necessary for really effective mixing.) A further simplification is to assume that the plain section can be represented by two plates with a single masterbatch striation between (Figure 2.15). A strain of 5 units gives an interfacial area increase of 5 approximately. For a large strain it would be the same as Cos ∅ → 1 38

Dispersive and Distributive Mixing

Figure 2.14 Laminar shearing interrupted by repeated cutting and turning

Figure 2.15 Laminar shearing interrupted by repeated turning: Details of first three stages. 39

Mixing in Single Screw Extrusion

Figure 2.16 Graph of mixing versus number of shearing-cutting-turning stages.

If we now cut the striation into 5 equal pieces, turn them at right angles to the row of pins, and we repeat the process, the area increase is now 25. If we had continued the strain without cutting and turning, it would have only increased to 10. If we now plot striation thickness ratio (or area ratio) between start to finish over the five rows of pins and compare with those without cutting and turning we get the graph shown in Figure 2.16. This concept was demonstrated by Ng and Erwin in experiments using coloured polyethylene sheared in an annulus formed between two concentric cylinders [21]. The procedure was to separately mould rings of black and white polyethylene in the annulus, cut these into segments, and replace them; alternating the black and white, with adjoining faces radial to the annulus (Figure 2.17). The apparatus was heated in an oven at 175° and the inner cylinder rotated to give an amount of shear strain (i.e., similar to Couette flow in Figure 2.8).

S=

40

R∅ W

(∅ = no. of radians turned)

Dispersive and Distributive Mixing

Figure 2.17 Erwin’s experiments using Couette flow with repeated orientations. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 2. ©1980, Rapra Technology)

After cooling to solidify, the sheared black and white annulus was cut into rectangular blocks and average striation thickness measured. The blocks were re-oriented and replaced such that the interfaces were perpendicular to the shearing ring. The melting, shearing, cooling, measurement of striation thickness and re-orienting were then repeated several times. Plots of Af /Ao (derived from striation thickness ratio ro/rf) against total shear strain, (known from the total amount the cylinder rotated) plotted on log scales gave close agreement to theory. Experiments were carried out by Bigio and co-workers [23] to evaluate mixing using five different screw geometries in an extruder in which a transparent barrel rotated around a fixed screw. An engineering lathe supplied both the barrel rotation and drive for two syringes which pumped black and white pigmented curable silicone as two separate streams into one end of the screw. On completion of extrusion, the silicone was cured within the screw and barrel assembly in an oven, and then unwound from the screw for sectioning and striation thickness measurement. 41

Mixing in Single Screw Extrusion

Figure 2.18 Influence of mixing pin.

Graphs were plotted for the number of striations (which is inversely proportional to striation thickness) against total average strain. These typically showed a straight line of constant slope. When a peg was introduced part way along the screw, the slope changed abruptly beyond the peg to a steeper gradient of about double the initial slope (Figure 2.18). The Maddock element (see Chapter 8) had a similar effect. It appears very likely that, had several equally spaced pins been used, a series of straight lines of increasing slopes would have resulted, possibly as in Figure 2.16.

2.3.4.2 Application of Repeated Re-orientations During the 1970s, most UK extrusion companies producing black polyethylene pipes and cables used pre-compounded material. If fed with a blend of natural polymer and masterbatch, which would have reduced material costs, they would have failed to meet the carbon black mixing requirements of the relevant British Standards. The only mixing device which appeared to be both suitable and available to them, was the Stanley (ICI) ‘turbine’ mixer [24, 25]. This system of alternating rows of fixed and moving teeth (see Chapter 9) appeared to be far superior to rows of teeth confined to the screw [26]. Unfortunately, this system sometimes caused operating problems (see Chapter 9). 42

Dispersive and Distributive Mixing The turbine system was followed by two alternative interacting mixers [27, 28], which had rotors and stators with overlapping cavities in the shape of Woodroffe key slots. These devices overcame the former’s disadvantages. These mixers are discussed in detail in Chapter 9. The author evaluated this type of configuration as a potential extrusion technique to achieve good mixing by reproducing Erwin’s mixing model [29]. The overall objective was to produce extrusions meeting black water pipe standards using carbon black masterbatch. Striation thickness would be measured using samples taken from cavity slots along the mixer. The mixing unit was made with a rotor attached to the screw tip of a 38 mm extruder turning within a stator fitted as an extension to the barrel flange. For experimental purposes, the rotor cavities were wide key slots arranged in rows, and the stator consisted of a matching sleeve having overlapping rectangular slots with semi-circular ends. The stator slots were closed to form cavities by sliding the sleeve into the closely fitting barrel (Figure 2.19). The stator bore was nominally the same as the extruder. An 8 mm strand die was fitted to the stator outlet. With the arrangement of in-line interchanging cavities, the rotor and stator could be removed as a unit with the screw. In the evaluation, clear plasticised PVC was extruded using this mixer with a thin streak of black plasticised PVC extruded through a transducer port situated at the adapter joining the mixer to the screw. The black streak was extruded from a 25 mm extruder. When conditions had stabilised, both extruders were stopped, the die removed and the screw jacked out of the barrel by a hydraulic ram fitted at the back of the extruder. The mixer stator, being an extension of the screw, moved out of the mixer housing taking the stator with it. The rotor was then pulled out of the stator sleeve, shearing the PVC at the cavity interfaces and the PVC from each stator cavity pushed out and labelled. Microscopy sections 20 µm thick were then microtomed from the specimens and negative prints with 10 times magnification were prepared by putting the mounted sections in the film carrier of a photographic enlarger (Figure 2.20). A graph of measured striation thickness against mixer stage is shown in Figure 2.21. In cavity 6, striation thickness was calculated from estimated number of striation lines per square on the microscope graticule, whilst in cavity 7, the striations could only be detected as lines of black specks and average striation thickness could not be measured. However, extrapolation of the graph gives a figure for cavity 7 of 0.01 µm, and as the recommended carbon black particle size for good UV protection is 0.01 to 0.02 µm, it would be of the same order as the striation thickness derived from extrapolation. 43

Mixing in Single Screw Extrusion

Figure 2.19 A2-B2 mixer. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 2.5. ©1980, Rapra Technology)

Figure 2.20 Photomicrographs from 1st, 4th, and 7th cavity row of A2-B2 mixer. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 5. ©1980, Rapra Technology) 44

Dispersive and Distributive Mixing

Figure 2.21 Striation thickness versus mixer stage for plasticised PVC with injected striation. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 7. ©1980, Rapra Technology)

Experiments were then carried out to compare specific output rate and masterbatch distribution. The material was natural low-density polyethylene blended with 5% of a 40% carbon black masterbatch. In Figure 2.22, a comparison of output rate against screw speed of the extruder without a mixer and with a slotted cavity mixer shows the performance to be virtually identical. Photomicrographs of microtomed sections of the extrudates produced without the mixer, contained the usual masterbatch streaks (Figure 2.23(a)), whereas those obtained from extrudates produced with the mixer had only very faint streaks (Figure 2.23(b)). A similar study was carried out by Edwards and Shales [30, 31], but in their experiments they used a mixer with a rotor and stator in which the parallel rows of slots were replaced with staggered rows of hemispherical cavities, as described in Chapter 9. The 32 mm extruder with an axially split barrel for opening and screw removal after cooling was fed with 2 mm pellets of high-density polyethylene (HDPE): 50 wt% black and 50 wt% white. Screw speed was 50 rpm and the die pressure was 16 MPa. Photomicrographs of sections removed from the screw channel 45

Mixing in Single Screw Extrusion

Figure 2.22 Output rate and melt temperature versus screw speed for 38 mm extruder fitted with the A2-B2 mixer. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 13. ©1980, Rapra Technology)

(a)

(b)

Figure 2.23 Photomicrographs from extrudates with and without A2-B2 mixer. (Reproduced with permission from G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Technology Members’ Report No.46, Rapra Technology, Shrewsbury, UK, 1980, Figure 10. ©1980, Rapra Technology)

showed the presence of striations, whilst microtomed sections taken from the mixer cavities showed a steady reduction in striation thickness until undetectable after passage through four cavity rows of the mixer (Figure 2.24). 46

Dispersive and Distributive Mixing

Figure 2.24 Photomicrographs showing improving mix quality over five cavity rows of a cavity transfer mixer using HDPE. (Adapted with permission from R.W. Shales, Mixing of Thermoplastics in Single Screw Rextruders, Department of Chemical Engineering, University of Bradford UK, 1989. [PhD thesis])

An investigation into the mixing mechanism within this type of mixer was carried out at Rapra [32], using a device with an aluminium rotor within a transparent acrylic stator having a similar overlapping cavity arrangement to that used by Edwards and Shales (See Figures 2.25, 2.26 and Chapter 9). The mixer was mounted vertically, with a room temperature curing transparent silicone elastomer pumped in at a controlled rate at the bottom, and a stream of the same elastomer containing pigment to provide a model striation, was injected at a constant rate through a hollow needle into a first row stator cavity. The unit had three cavity rows in the stator overlapping two complete and two half cavity rows at entry and exit in the rotor. Each row had six cavities equally spaced circumferentially. Observations were recorded on still and video cameras, and after overnight curing of the silicone elastomer, the mixer was dismantled and the silicone castings were sectioned and flow patterns examined. 47

Mixing in Single Screw Extrusion

Figure 2.25 Model Cavity Transfer Mixer with transparent acrylic stator. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.1. © 1984, Rapra Technology)

Colour injection at different depths showed that laminar flow took place which was generally similar to that occurring in the channel of an extruder screw. A globule injected from a hypodermic syringe was quickly transformed into a striation when the rotor turned through a comparatively small angle. The continuously injected striation travelled around the cavity in the direction shown in Figure 2.27. Progression through the mixer is represented diagrammatically in Figure 2.28, which depicts the cross sectional plane of part of the stator with the rotor cavities passing in an anti-clockwise direction. The lands between the cavities are numbered round in a clockwise direction. Both true cross-sectional representation and flow paths are very complex, but this simplified picture gives a good representation of the overall pattern. 48

Dispersive and Distributive Mixing

Figure 2.26 Model Cavity Transfer Mixer with half acrylic stator removed to show aluminium rotor. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.2. © 1984, Rapra Technology)

Figure 2.27 Flow streamlines in a stator cavity. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.3. © 1984, Rapra Technology) 49

Mixing in Single Screw Extrusion

Figure 2.28 Diagrammatic representation of mixing action. (Reproduced with permission G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.8. © 1984, Rapra Technology)

50

Dispersive and Distributive Mixing

Figure 2.29 Formation of multiple striations. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.9. © 1984, Rapra Technology)

The injected striation travels in a clockwise direction around the surface of the cavity and then crosses over and turns sharply in the direction of rotation of the rotor. It is then carried around inside the rotor, being influenced by laminar flow within the rotor cavity and stator cavities and lands which it passes. It is then cut off into small segments and deposited in stacked layers in a stator cavity of the next row along. Although represented in two dimensions as striations, they are by then broad ribbons. Following the cutting and transfer actions, the striations are now at right angles to the laminar shear direction, the centre of the striation moving at right angles with respect to the ends, so that the striation (or ribbon) becomes folded. The repetitive action produces a stacking effect, such that packed groups are then subjected to laminar shear, cutting and transfer as for the original single striation. This progression is depicted in Figures 2.29 and 2.30 which shows stator cavities in three successive rows. Figure 2.31 is a photomicrograph of a sectioned casting showing striations leaving cavity two for a rotational speed of 5.5 rpm, where the stacking and shearing at right 51

Mixing in Single Screw Extrusion

Figure 2.30 Progress of striation through mixer. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.10. © 1984, Rapra Technology)

angles can be clearly seen. A further contribution to mixing is the folding back of the cut ends of the striations by the passing land which produces a pattern of ‘hooks’ at the ends of the ‘stacked ribbons’. Similar patterns were found in cavity 3 for a number of rotational speeds. Although the CTM is geometrically very simple from a machining aspect, the complex geometry of the moving boundaries makes a detailed analysis of the flow field a difficult task according to Wang and Manas-Zloczower [33]. In this paper (following seven earlier ones) [34-40], a model CTM was used with silicone fluid viewed through a transparent window. Particle tracking of the 0.1% carbon black added to the silicone flow simulation media was made with a camcorder. The computer simulation showed very good agreement with the experimental results. Figure 2.32 shows Shear stress and elongational flow component versus cavity depth. Reducing cavity depth increased shear stresses, more so for the stator than the rotor. Y, the 52

Dispersive and Distributive Mixing

Figure 2.31 Striations leaving cavity No.2 at 2 rpm. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 8.12. © 1984, Rapra Technology)

Figure 2.32 Average shear stresses and y values for different designs. (Reproduced with permission from C.C. Wang and I. Manas-Zloczower, International Polymer Processing, 1996, 11, 2, 115, Figure 7. ©1996, Carl Hanser Verlag) 53

Mixing in Single Screw Extrusion elongational flow component, ranged between 0 for pure rotation and 1 for pure elongation. It appeared that shallower cavities on both rotor and stator had better potential for dispersive mixing. To investigate dispersive mixing, 2 mm spheres of ‘fluffy’ carbon black were moulded into ‘model agglomerates’, densities 0.3 and 0.4 g/cm3 and a device (#2) with 10 mm cavities in both rotor and stator compared with device #1 (deep cavities). Neither were dispersed in #1 (deep cavities), nor was 0.4 in #2 (shallow cavities). However, the 0.3 density agglomerate was dispersed in #2 by tail formation and erosion as described in Chapter 14. Earlier theoretical treatments of this type of mixer have been made by Lin and Bevis [41, 42], de Jong [43], and Bromilow and Hulme [44]. The overall picture is that the repeated shearing and repositioning required by the theory appears to be present within the repeating complex patterns as the hemispherical cavities, (together with a network of lands), pass each other. An interesting feature (as demonstrated in Chapter 8), is that larger but fewer cavities will mix as well as smaller but more numerous cavities. The former arrangement minimises pressure drops across the mixer.

2.3.4.3 Summary of Mixing Achievable in a Single Screw Extruder The statements listed next summarise the fundamentally limited distributive mixing of a conventional extruder and the achievable advanced level required to meet specific technical and aesthetic requirements at economic output rates: 1) Shear stresses will be comparatively low, and hence, it follows that masterbatches will be necessary for good dispersive mixing of pigments, flame retardants and fillers and so on. 2) Laminar shear mixing will occur which will provide distributive mixing. 3) Although such mixing is available from the point of melting to the screw tip, orientation of striations formed from melted pellets may well result in these striations persisting through the extruder and into the product. 4) Interrupting the laminar shear mixing with regular turning, ideally at right angles, will achieve good distributive mixing. 5) Mixing devices such as peg mixers will contribute to this mixing requirement, and may be adequate for many products. 54

Dispersive and Distributive Mixing 6) For complete elimination of striations (as may be required for polyethylene pipes and cables), the required mixing mechanism can be achieved by using interacting stationary and moving parts. 7) The mechanisms can be alternating fixed and moving teeth, or overlapping fixed and moving cavities. Such systems will be more expensive than most screw mixing elements, but overall will be justified by reduced extrusion costs for products where very good mixing is necessary.

References 1.

W.R. Bolen and R.E. Colwell, SPE Technical Papers, 1958, 4, 98, 1004.

2.

P.K. Freakley and W.Y. Wan Idris, Rubber, Chemistry and Technology, 1979, 52, 134.

3.

G.M. Gale, inventor; no assignee; GB 2,422,327, 2003.

4.

T. Theodorou, Department of Chemical Engineering and Chemical Technology, Imperial College, London, UK, 1978. [MSc thesis]

5.

G.T. Taylor, Proceedings of the Royal Society, Series A, 1934, 146, 501.

6.

J.J. Elmendorp in Mixing in Polymer Processing, Ed., C. Rauwendaal, Marcel Dekker, New York, NY, USA, 1991.

7.

S.V. Kao and S.G. Mason, Nature, 1975, 253, 5493, 619.

8.

M.J. Smith, Journal of the Oil Colour Chemists’ Association, 1973, 56, 3, 126.

9.

M.J. Smith, Journal of the Oil Colour Chemists’ Association, 1973, 56, 4, 155.

10. M.J. Smith, Journal of the Oil Colour Chemists’ Association, 1973, 56, 5, 165. 11. M.J. Smith, Journal of the Oil Colour Chemists’ Association, 1974, 57, 1, 36. 12. V.G. Ammons, Industrial and Engineering Chemistry, 1963, 55, 40. 13. H-H. Bunge, Kunststoffe, 1986, 73, 12, 1214. 14. H. Schlumpf, Kunststoffe, 1983, 73, 9, 511. 15. F.N. Cogswell, Polymer Melt Rheology, George Godwin, London, UK, 1981. 55

Mixing in Single Screw Extrusion 16. J.M. McKelvey, Polymer Processing, John Wiley & Sons, London, UK, 1962, Chapter 12. 17. Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley & Sons, New York, NY, USA, 1979, Chapter 12. 18. J. Smith in Proceedings of a Plastics and Rubber Institute Conference Polymer Extrusion 2, London, UK, 1982, Paper No.20. 19. L. Erwin in Mixing in Polymer Processing, Ed., C. Rauwendaal, Marcel Dekker, New York, NY, USA, 1991, Chapter 1. 20. L. Erwin, Polymer Engineering and Science, 1978, 18, 7, 572. 21. K.Y. Ng and L. Erwin, Polymer Engineering and Science, 1981, 21, 4, 212. 22. R.S. Spencer and P.M. Wiley, Journal of Colloid Science, 1951, 6, 133. 23. D.I. Bigio, J.D. Boyd, L. Erwin and D.W. Gailus, Polymer Engineering and Science, 1985, 25, 5, 305. 24. T.A. Stanley, inventor; ICI Ltd., assignee; GB 787,764, 1955. 25. T.A. Stanley, inventor; ICI Ltd., assignee; GB 843,849, 1957. 26. G. Martin, Kunststofftechnik, 1972, 11, 12. 27. K. Gerber, inventor; Metal Box, assignee; US 3,174,185, 1962. 28. P. Renk, inventor; Barmag, assignee; US 4,253,771, 1978. 29. G.M. Gale, Distributive Mixing in Plastics Extrusion, Rapra Members Report No.46, Rapra Technology, Shawbury, Shrewsbury, UK, 1980, p.3. 30. M.F. Edwards and R.W. Shales in Proceedings of a Rapra Technology Ltd, Conference - Making the Most of the Cavity Transfer Mixer, Shawbury, Shrewsbury, UK, 1986, Paper No.2. 31. R.W. Shales, Mixing of Thermoplastics in Single Screw Extruders, Department of Chemical Engineering, University of Bradford, UK, 1989. [PhD thesis] 32. G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984. 56

Dispersive and Distributive Mixing 33. C.C. Wang and I. Manas-Zloczower, International Polymer Processing, 1996, 11, 2, 115. 34. C.C. Wang and I. Manas-Zloczower, International Polymer Processing, 1994, 9, 1, 146. 35. C.C. Wang and I. Manas-Zloczower, Polymer Engineering and Science, 1994, 34, 1224. 36. J.J. Cheng and I. Manas-Zloczower, Polymer Engineering and Science, 1989, 29, 701 37. J.J. Cheng and I. Manas-Zloczower, Polymer Engineering and Science, 1989, 29, 1059. 38. J.J. Cheng and I. Manas-Zloczower, Journal of Applied Polymer Engineering Science: Applied Polymer Symposium, 1989, 44, 35. 39. H.H. Yang and I. Manas-Zloczower, International Polymer Process, 1992, 7, 195. 40. H.H. Yang and I. Manas-Zloczower, Polymer Engineering Science, 1992, 32, 1411. 41. S.Y. Lin and M.J. Bevis, Plastics Rubber and Processing Applications, 1987, 7, 1, 29. 42. S.Y. Lin, Plastics and, Rubber Processing and Applications, 1987, 8, 133. 43. E. de Jong in Proceedings of a Rapra Technology Ltd., Seminar - Making the Most of the Cavity Transfer Mixer, Shawbury, Shrewsbury, UK, 1988, p.14. 44. T.M. Bromilow and A.T. Hulme, Proceedings of a Rapra Technology Ltd., Seminar - Making the Most of the Cavity Transfer Mixer, Shawbury, Shrewsbury, UK, 1986, Paper No.3.

57

Mixing in Single Screw Extrusion

58

3

Measurement of Mixing

3.1 The Need for Measurement of Mixing As explained previously, in Chapter 2, mixing can be divided into dispersive and distributive mixing. The quality of the supplied compound or masterbatch depends on the former with the extrudate quality dependent on the latter being achieved to the required degree during single screw extrusion of the final product. Dispersive mixing problems are largely manifested by the presence of agglomerated pigment, filler, and other solid additive particles, with distributive mixing problems producing striations. Although materials suppliers will carry out their own quality tests, it is normal for cable and pipe producers to carry out mixing quality testing on their products to the relevant standards as single screw extruders have inherently limited distributive mixing limitations as well as being generally poor dispersive mixers. Replacement of failed cables and pipes can incur very high costs. In addition to pipe and cable producers, extruders of many other products should be concerned with the quality of distributive mixing of masterbatch as this may well effect many aspects of technical performance, appearance and cost. The fire test failure described in Chapter 1 is a good example. In product extrusion it may also be necessary to assess how much better or worse mixing has become as a result of making changes to the operating conditions or to the machine itself. For example, raising screw speed to increase output rate may delay completion of melting such that the mixing efficiency is impaired to the point where the extrudate is unacceptable. Laboratory studies may require the progress of mixing through the various stages of the extruder to be followed such that measurement of mixing achieved at a number of points will be necessary. Examination on a fine enough scale will observe the mixture components. These are most likely to exist either in the form of striations or as gritty particles. Lumpy and uneven surfaces may also be observed. In most cases, striations will be the observed criteria, but if dispersive mixing is the limitation, particles over a particular size and/ or exceeding a specified concentration will determine the outcome. 59

Mixing in Single Screw Extrusion The criteria might be visual appearance, microscopic appearance or measured properties. Examples of all these were included in Chapter 1. As mentioned in Chapter 2, 10-4 m is approximately the resolution of the human eye [1], but other criteria may apply, for example, resistance to weathering, fire resistance, surface friction, anti-block and so on.

3.2 Striation Thickness Measurement Mixing during single screw extrusion produces layered structures, and the interfacial area between the two components is a measure of the degree of mixing. This would be difficult to measure but for the formation of striations from masterbatches and mixtures of coloured pellets. With their progressive thinning during passage through the extruder until ideally they are no longer readily observable under a microscope, measurement of striation thickness is a practical alternative. It can be used for both following stages of mixing of samples from the screw channel and assessing mixing achieved in the extrudate. For such practical measurements, striation thickness is directly related to interfacial area [2], the simple relationship being:

Ao =

1 S

S is defined as one half of the repeat unit shown in Figure 3.1.

Figure 3.1 Relationship between striation thickness and interfacial area.

Striation thickness can be measured using microtomed sections examined under a microscope using transmitted light. The thickness can be measured manually using a 60

Measurement of Mixing microscope with a graticule eye piece or by using enlarged photomicrographs. This system is sufficiently accurate to measure comparatively large changes in striation thickness, but is inadequate to quantify small changes or deal with wide variations. For example, the microtomed sample of an extrudate (Figure 11.4), produced using a static mixer, would involve a long and tedious operation to produce a representative striation thickness from this photomicrograph. Image analysis is a less judgemental process for producing a representative value. Benkreira and co-workers [3] have described the method they used on samples removed from the extruder screw channel after rapid cooling to freeze the molten polymer, mainly at the stage where the polymer had just melted. Samples cut from the screw helix were microtomed in sections 20 µm thick and examined under a 100x magnification microscope using transmitted light. Image analysis was then used to detect grey levels over a range of 256 ranging from completely black to completely white. The system picked out the same or darker grey levels than a chosen level. The data was processed via a microcomputer interface into indices assessing mixing, in this case striation thickness of the minor component.

3.3 Agglomerate Measurement As discussed in the previous chapters, single screw extruders are not good dispersive mixers, and therefore reliance is placed on suppliers to provide well dispersed additives in compounds and masterbatches. When it is considered necessary to test incoming materials or compare samples from different suppliers, which may, in particular, be required for carbon black masterbatch, there are a number of quality standards. These are mainly for carbon black in polyethylenes used in water pipes and cables, but can also be used for coloured pipes and cables where agglomerates can cause electrical failures. The tests can be divided between those that examine thin samples, and those which use extrusion filtering.

3.3.1 Microscopy Examination of Thin Samples The most common visual assessment for carbon black is made by comparing thin samples under a microscope with a set of photomicrographs in the standard [4]. This requires matching the specimens’ appearance with those in the standards, as a specified number of agglomerates under a certain size is permissible The specimen can be a microtomed section or a tiny piece pressed out between glass microscope slides on a hot plate. The hot pressing (sometimes referred to as the ‘pinhead’ test) tends to smear striations, but is suitable for detecting agglomerates. 61

Mixing in Single Screw Extrusion

3.3.2 Agglomerate Count for Blown Film Both compounds and masterbatches containing carbon black or pigments, can be assessed for agglomerates using film samples prepared by extrusion film blowing. The extruded materials can be compounds or masterbatches diluted with natural polymer. Assessment is by laying a film sample on a light box and counting the number of visible agglomerates per unit area. For carbon black, a final concentration of 1% masterbatch has been recommended [5]. Agglomerate counts by image analysis are possible on microtomed sections and on polymer streams passing a window [6].

3.3.3 Screen Pack Filtration Test An alternative approach is to determine the extent of the presence of agglomerates by extrusion filtration. There are two techniques based on extrusion of a fixed quantity of compound or masterbatch through a screen pack. This eliminates the very small scale of examination and human judgement involved in the microscopy tests, which may also view unrepresentative areas. Measurement is by one of two techniques: 1) A microscope count of the number of agglomerates retained on the wire mesh screen pack. 2) Measurement of pressure build-up due to clogging of the mesh [7].

3.3.3.1 Agglomerate Retention on a Wire Mesh Screen This can be illustrated by describing comparisons made for four compositions containing low-density polyethylene (LDPE) and carbon black as used at the time for pipes meeting standards for cold water services [8]. These were as follows: 1) A compound of commercial origin supplied as satisfactory. 2) A compound of commercial origin rejected by the supplier as unsatisfactory but supplied as a sample for comparison. 3) A pellet blend of natural LDPE and 5 wt% of a carbon black masterbatch. 4) A particulate blend of natural LDPE and 2 wt% carbon black powder. Comparisons were made to BS3412:1976 (superseded by BS 3412:1992 [9]) using a 25 mm extruder equipped with a 150 µm aperture screen supported by 250 µm, 420 µm screens and a breaker plate. After extruding 640 g ± 10 g and purging the 62

Measurement of Mixing black from the machine with natural polymer, the strand die and screen/breaker plate assembly were removed. The number of agglomerates retained on the 150 µm screen was counted using an 8× magnification microscope. The standard required there to be no more than 70. The results are shown in Table 3.1, and photomicrographs in Figures 3.2 to 3.8. The screens for the satisfactory and unsatisfactory compounds, and diluted masterbatch are shown in Figures 3.2–3.4 and clearly illustrate the differences recorded in Table 3.1.

Table 3.1 Screen pack extrusion test for agglomerates on four materials Polymer Composition

Screw speed Motor Melt temperature (rpm) current (amp) (°C)

Agglomerate count

Unsatisfactory compound

40

5.0

213

120-140

Satisfactory compound

40

5.0

216

23

Masterbatch

40

4.0

219

91

Dry blend

40

4.0

215

Not measurable

Figure 3.2 Filter screen showing agglomerates for an unsatisfactory compound. (Reproduced with permission from M. Penny, Investigation of a Roller Bearing Mixer for Extrusion Compounding of Carbon Black with Polyolefines, Rapra Members Report No. 45. Rapra Technology, Shawbury, Shrewsbury, UK, 1980, Figure 3.1. ©1980, Rapra Technology) 63

Mixing in Single Screw Extrusion

Figure 3.3 Filter screen showing agglomerates for a satisfactory compound. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.2. ©1980, Rapra Technology)

Figure 3.4 Filter screen showing agglomerates for a nearly satisfactory masterbatch/natural polymer pellet blend. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.4. ©1980, Rapra Technology)

64

Measurement of Mixing

Figure 3.5 Cross section of a strand extruded from a satisfactory compound. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.5. ©1980, Rapra Technology)

Figure 3.6 Cross section of a strand extruded from a masterbatch/natural polymer blend. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.6. ©1980, Rapra Technology)

65

Mixing in Single Screw Extrusion

Figure 3.7 Cross section of a strand extruded from a carbon black/natural polymer dry blend. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.6. ©1980, Rapra Technology)

Figure 3.8 Satisfactory compound section based on a microtomed x100 photomicrograph. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.8. ©1980, Rapra Technology) 66

Measurement of Mixing The photomicrographs of the cross sections of extruded strands produced without the screen-pack are shown in Figures 3.5-3.7. Figure 3.5 is for the satisfactory compound. Figure 3.6 is the one containing masterbatch, and Figure 3.7 the carbon black plus virgin polymer dry-blend. They illustrate the very different results from the three different methods of incorporating carbon black. The satisfactory compound produced a strand with a uniform appearance, whilst the masterbatch plus virgin polymer blend had numerous striations, demonstrating the lack of distributive mixing. The dry-blend with its numerous agglomerates and striations clearly demonstrated a complete lack of dispersive and distributive mixing as expected. Figures 3.8 and Figure 3.9 have been reproduced from the original ×100 magnification photomicrographs prepared according to the British Standards test. Figure 3.8 shows a solitary small agglomerate and no striations for the satisfactory compound, whilst Figure 3.9 shows the masterbatch plus virgin polymer blend has unpigmented striations, but no agglomerates. However, one or two may be hidden within the black layers.

Figure 3.9 Masterbatch natural polymer blend section based on a microtomed ×100 photomicrograph. (Reproduced with permission from M. Penny, Roller Bearing Mixer for Extrusion of Carbon Black with Polyolefines, Rapra Members Report No.45, Rapra Technology, Shawbury, Shrewsbury, UK, Figure 3.10. ©1980, Rapra Technology) 67

Mixing in Single Screw Extrusion 3.3.3.2 Pressure Build-up Caused by Agglomerate Retention on a Wire Mesh Screen (Filter-Pressure Value) During the straining of agglomerates from masterbatches and compounds using a screen for subsequent visual examination and agglomerate counts under a microscope, the melt pressure at the barrel exit will steadily rise as a result of screen blockage by the agglomerates. Consequently, the recorded melt pressure after extrusion of a fixed amount of material will be a measure of the quality of dispersion of the carbon black or pigment. This may be a useful method for re-assurance that mix quality with regard to dispersion equals or exceeds a level known to be representative for that particular product. An advantage is the elimination of judgmental microscope counts, plus an overall time saving. A disadvantage is that the increasing melt pressures may give an improved result. To eliminate this problem, DIN-EN 13900-5 [10] was issued in which a gear pump was introduced between the extruder and screen pack to provide a consistent back pressure for the extruder. This filter-pressure value test starts and finishes with extrusion of the matrix polymer without masterbatch [7].

3.4 Influences of Mixing on Product Properties The incorporation of additives is carried out to achieve a technical specification, some of which will be more sensitive to extruder mixing performance than others. Technical specifications are normally covered by standards testing, whether to recognised standards or in-house specifications. A concern here is to recognise the extruder’s mixing contribution to achieving, or (more likely), the failure to meet standards such as fire retardancy and impact tests described in Chapter 1. Where agglomerates appear to be causing, for example, impact and electrical failures, the first suspect is the masterbatch or compound supplier. As a first step, it may be advisable to check whether the undispersed specks are one of the following: 1) Dirt or other contaminants such as packing material, incompatible polymeric ‘bits’ from scrap granulation. (Even dandruff has been found in optically clear film). Burnt and degraded polymers 2) Gels due to oxidation during extrusion, possibly caused by long coextrusion feed pipes and dead corners in film dies (see Chapter 14). Pigmented gels can easily be mistaken for agglomerates. (Note that it can often save time and overall costs by having unrecognised particles chemically analysed. Infra red spectroscopy can often identify very small specks and can also discriminate between carbon black agglomerates and black polymer gels.) 68

Measurement of Mixing

3.5 Preparation of Thin Sections for Optical Microscopy Assessment Although polymer blending papers often include a description of the preparation of samples for electron microscopy, little is published on the preparation of samples for optical microscopy. Optical microscopy has sufficient magnification to measure the width of thin masterbatch striations, agglomerate sizes described in pipe and cable standards etc. It also has an advantage in that comparatively large areas can be viewed as illustrated in the pipe samples in Chapter 1, Although these particular pictures required considerable experience and care, the same general techniques were learned and applied by a number of University students on industrial placements at Rapra who were involved in much of the practical extrusion work on mixing described in Chapters 2, 3, 7, 8 and 9. For these reasons, the techniques used (which were developed by Ivan James) have been included as an Appendix at the end of the book.

References 1.

J. Smith in Proceedings of a Plastics and Rubber Institute Conference Polymer Extrusion 2, London, UK, 1982, Paper No.20.

2.

J.M. McKelvey, Polymer Processing, John Wiley & Sons, London, UK, 1962, Chapter 12.

3.

H. Benkreira, R.W. Shales and M.F. Edwards, International Polymer Processing, 1992, 7, 2, 126.

4.

BS ISO 18553, Methods for the Assessment of Pigment or Carbon Black Dispersion in Polyolefin Pipes, Fittings and Compounds, 2002.

5.

Carbon Black Dispersion, Technical Report S-131, Cabot Corporation, Billerica, MA, USA, 1989.

6.

D.W. Yu, M. Esseghir and C.G. Gogos in Proceedings of the Annual SPE Conference – ANTEC ‘96, Indianapolis, IN, USA, 1996, p.136.

7.

J. Schut, Plastics Technology, 2005, 51, 7, 45.

8.

M. Penny, Investigation of a Roller Bearing Mixer for Extrusion Compounding of Carbon Black with Polyolefines, Rapra Members Report No. 45. Rapra Technology, Shawbury, Shrewsbury, UK, 1980. 69

Mixing in Single Screw Extrusion 9.

BS 3412:1992, Methods of Specifying General Purpose Polyethylene Materials for Moulding and Extrusion, 1992.

10. DIN EN 13900-5, Pigments and Extenders – Methods of Dispersion and Assessment of Dispersibility in Plastics – Part 5: Determination by Filter Pressure Value Test, 2005.

70

4

Single Screw Extruder Stages: Effects on Mixing

The continuous progress of plastics materials from hopper to die involves a number of stages, all of which can influence how well mixed the extrusion will be. These various stages form the backbone of Chapter 4 to Chapter 8 in which the polymer’s journey is followed through the extruder. Such a journey has been described by Tadmor and Gogos [1]. This covers the influence of each stage on mixing whilst extra devices which can be added to the end of the extruder to raise the overall level of mixing are included in Chapters 9-11. Figures 4.1, 4.2 and 4.3 provide a ‘roadmap’ for this journey, and ties together Chapters 5 to 11. 1) Figure 4.1 is a diagram of a conventional single screw extruder and shows the following stages: hopper feed, pellet conveying, compression/melting, metering and pumping. 2) There is an unrolled screw channel and a block diagram of the same extruder, showing the stages and actions performed by the extruder. These are particularly relevant to the extrusion of products in a situation where unacceptable laminar striations due to inadequate mixing, occur in the product. The overall situation in this figure is: • There is no mixing in the feed zone • Melting is a prerequisite for mixing to occur • Un-melted material can break away and float downstream in the melt such that it tends to flow down channel in an un-melted state and may remain unmixed. • Laminar shear flow in the melt zones of the screw may be inadequate to prevent striations • Late completion of melting gives less opportunity for mixing 71

Figure 4.1 Stages affecting melting and mixing: conventional extruder.

Mixing in Single Screw Extrusion

72

Figure 4.2 Stages affecting melting and mixing: extruder with barrier screw plus shearing and mixing elements.

Single Screw Extruder Stages: Effects on Mixing

73

Figure 4.3 Guide to chapter numbers for extruder functions affecting mixing.

Mixing in Single Screw Extrusion

74

Single Screw Extruder Stages: Effects on Mixing Figure 4.2 provides a guide to various operations and associated components which can be used to reduce or prevent the mixing shortcomings illustrated in Figures 4.1. It shows the extruder again, but with the following additions: 1) A barrier flight screw to control rate of melting. 2) A barrier element to ensure no un-melted polymer goes any further than this point. It will also provide some distributive mixing. 3) A distributive mixing device. In this example, several rows of pins. The unrolled screw channel and block diagram of this extruder shows the stages and actions performed by the extruder with the added components. Additional to the controlled melting, is the replacement of the slow laminar shear mixing with repeated re-positioning and laminar shearing by the pin mixer. The range and effectiveness of such mixing devices is very wide and described in Chapter 8. Figure 4.2 also shows the additional mechanisms featured in the block diagram. It should be noted that any one, two or all three of these features might be used. Although it could be argued that with an efficient melting screw, the additional melting device would be unnecessary, this combination appears to be widely used. Figure 4.3 provides a guide to the following chapters and also refers to the less widely used but effective mixing devices (covered by Chapters 9, 10 and 11) that are considered specialist items ideally suited for specific applications.

References 1.

Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley & Sons, New York, NY, USA, 1979.

75

Mixing in Single Screw Extrusion

76

5

Pellet Handling: A Source of Variable Composition

5.1 Introduction The particulate properties of polymers, blends and masterbatches can affect any or all of the following operations: • Bulk handling, including silo emptying • Auger metering • Particulate blending • Hopper flow • Screw channel filling • Screw conveying:

1) In barrels with smooth feed zones



2) In barrels with grooved feed zones

For potential problems and solutions in bulk handling and storage of plastics materials, see the book on this subject by Butters [1]. Although no incorporation of additives in the polymer will occur until melting commences in the screw channel, it is useful to appreciate that additive concentration may vary due to behaviour of the feed materials. Such changes can cause product non-conformity ranging from changes in colour to varying technical performance such as flame retardancy. The basic problem is that the feed material can be a blend of several materials in which the particulate form consists of a range of sizes, shapes, densities and surface friction. As a result, no matter how well the particulate blending operations has been performed, the constituents will separate under certain unfavourable conditions. This is not surprising when considering that masterbatch pellets can cover a range of densities depending on the quantity and specific gravity of the additive, e.g., barium sulfate (density of 4.5 g/cm3) compared with typical polyolefine densities of about 0.9-0.96. 77

Mixing in Single Screw Extrusion Furthermore, the surface properties will be influenced by additives ranging from slippery waxes and slip additive for packaging films, to sticky additives for pallet and silage wrap films. The shape and size of the particulates may be a mixture of strand cut and die face cut pellets, with irregular granulated platelets of edge trim and skeletal scrap from thermoforming. Incorporation of up to 40% of the latter may be necessary to meet packaging thermoforming economics.

5.2 Hopper Design Whatever handling systems are used, some sort of hopper is necessary to feed polymer pellets into an extruder and in one direction at least, the final dimension is limited to the screw diameter. The other variables are the length of screw exposed, and the cross section, height and angle of the hopper itself. Figures 5.1-5.4 show a number of variations. The angle of the sides defines whether the hopper is a ‘mass flow’ or a ‘non-mass flow’ hopper. An angle greater than approximately 60° will normally give ‘mass flow’, and less will give ‘non-mass flow’, but both features can be combined, e.g., two opposite sides may have completely different angles.

5.2.1 Mass Flow Hopper This design (Figure 5.1) is characterised by ‘first in will be first out’. The surface of the particulate materials in the hopper will be approximately level and fall at the same rate from centre to the sides. Pressure at the outlet will depend on the height of material in the hopper, i.e., like water in a tank. Consequently, output rate may vary if the height of material has wide variations with time, e.g., 25 kg of pellets are tipped in from a bag when the hopper is almost empty.

Figure 5.1 Mass flow hopper. 78

Pellet Handling: A source of Variable Composition As a hopper containing comparatively large amounts of material will need to be high; it may require extra support and it will be harder to load from ground level.

5.2.2 Non-mass Flow Hopper With sides less than approximately 60°, these hoppers can be much lower and the pressure at the outlet will be less dependent on the height of the material (Figure 5.2). The material flows from the sides to the centre. As the material level drops in the centre, pellets rolling inwards down the slopes can easily segregate depending on density, shape and possibly friction coefficient, so that they collect in higher concentrations at the centre. Sticky materials, very irregular shaped material and some powders may result in ‘rat-holing’ in which material flows only from the centre leaving a vertical tubular hole with the bulk left stuck to the hopper wall (Figure 5.3).

Figure 5.2 Non-mass flow figure.

Figure 5.3 ‘Ratholing’. 79

Mixing in Single Screw Extrusion

5.2.3 Round Hoppers Round hoppers (Figure 5.4) look neat and are possibly of slightly lower cost. They are less likely to have ledges and corners and will be easier to clean. A round hole in the extruder barrel makes the hopper easier and cheaper to fit. However, as a round hole is limited to the screw diameter, it may restrict pellet discharge and encourage bridging with sticky or flaky materials. The inverted cone shape and relatively small round outlet may cause bridging due to formation of a stable arch, which may collapse and reform, causing extruder output variations. However, being round, they can more easily accommodate a stirrer to break up any sticking or bridging particulates, but this may reduce or increase risks of particle segregation, depending on the nature of the feed materials (Figure 5.5).

Figure 5.4 Round hopper.

Figure 5.5 Stirrer. 80

Pellet Handling: A source of Variable Composition

5.2.4 Square and Rectangular Hoppers As the discharge opening of square or round hoppers (Figure 5.6) is no longer round, and the length can be greater than the screw diameter, pellets will pass into the screw channel much more freely with less opportunity for bridging. A rectangular extruder opening can have various geometries to encourage channel filling and conveying, e.g., a tangential entry.

Figure 5.6 Rectangular hopper.

5.2.5 Ledges and Corners Joints between hopper and extruder and also in a sliding or swivelling ‘cut-off’ can result in ledges and corners where pellets may collect and fall out at some future date (Figure 5.7). This is possibly less likely with round hoppers.

Figure 5.7 Effect of ledge. 81

Mixing in Single Screw Extrusion

5.3 Composition Variations The most serious is probably the separation of additive pellets during flow from side to centre in a non-mass flow hopper.

5.3.1 Example 1 Following development of a pilot production line for extrusion of optical fibre ducting using a halogen-free flame retarded polyolefin compound, orders were placed in which several identification colours were needed, similar to insulated wires. Colour masterbatches were pre-blended with the unpigmented compound by batch tumbling weighed amounts. The line ran satisfactorily with several colours until an orange masterbatch was used. The ducting colour became very variable and it was found that the colour masterbatch pellets were separating from the main mix. (A ‘non-mass flow’ hopper was being used). The solution was to add the masterbatch from a small volumetric doser into a length of copper pipe fitted inside the hopper with a small funnel at the top and terminating just above the extruder screw (Figure 5.8).

5.3.2 Example 2 Blow moulded bottles were pigmented by pre-blending natural polymer with a yellow masterbatch in steel drums using an end-over-end blender. By dedicating individual drums to a particular colour, potential cross-contamination of colours was minimised, and cleaning was unnecessary. Transfer was by a standard vacuum hopper loader to the blow moulder hopper. The yellowness of the bottles varied over time, becoming very intense for short periods. It was found that the steel drum, with the vacuum loader pipe pushed down inside the pellets, behaved for periods as a non-mass flow hopper with yellow masterbatch pellets separating around the pipe entry until a point was reached where an accumulation of yellow masterbatch pellets was loaded. The answer was to vacuum load the natural pellets into a proprietary dosing/mixing device mounted on the extruder.

5.3.3 Example 3 After extruding black polyethylene pipe for a period using natural high-density polyethylene with carbon black masterbatch, the line was switched to extruding blue pipe using pre-compounded material. It took quite a long time to completely purge the extruder and die of black polyethylene, but eventually it was satisfactory. However, over a period of about half an hour, an occasional long black striation appeared in the blue pipe, before becoming completely uniformly blue again. It was evident that 82

Pellet Handling: A source of Variable Composition

Figure 5.8 Avoidance of pellet separation in hopper.

Figure 5.9 Masterbatch side feeder at hopper base.

the black masterbatch pellets had lodged at joints in the hopper, or possibly around bolts holding the sight glass and emptying side shute. The pellets had eventually dislodged and contaminated the blue pipe.

5.3.4 Other Systems Various arrangements are possible to deal with particular situations: one of the many possibilities is shown in Figure 5.9. 83

Mixing in Single Screw Extrusion

5.4 Measurement of Particulate Properties With the widespread use of automatic blending and hopper loading, as well as the range of hopper shapes used, the composition of the feed material may change due to particulate segregation following change of material supplier or even machine changes by the same supplier. Thus, sub-standard extrusions can arise following the un-noticed change from strand-cut to die face cut incoming material. Two pertinent points can be considered: 1) Materials which have good hopper flow may well have poor screw conveying, and vice versa. 2) There is a general lack of standards and readily available testing equipment suitable for testing plastics pellets and other particulate forms such as flake, crumb, shredded film, etc. There are both general standards and specific standards for such properties as apparent density, bulk factor and pourability of plastic materials, e.g., ASTM D1895 [2, 3]. Although the various standard funnel tests appear inappropriate for measuring the flow behaviour of plastics pellets and granulated scrap in extruder hoppers, a test using a range of funnels described by Boysen and Gronebaum [4], might be suitable. This method establishes the smallest of 7 funnel outlets through which the particulates will flow. Although aimed at testing polyvinylchloride powder compounds, the same principle could be used for pellets.

5.4.1 Hopper Flow Tests Testing standards involving timed discharges have little relevance to extruder hoppers. Under normal extrusion conditions, the slowly moving pellets lack the momentum and entrained air assistance associated with the standard tests. The test rig shown in Figure 5.10 is near to the real extruder situation. Controlling the discharge rate with a pair of rotating sponge rollers, avoids the assistance of momentum and air entrainment effects of standard tests, whilst using a redundant extruder hopper or a copy will replicate behaviour in a production extruder. 84

Pellet Handling: A source of Variable Composition

Figure 5.10 Hopper flow test using controlled discharge rate.

References 1.

G. Butters, Plastics Pneumatic Conveying and Bulk Storage, Applied Science Publishers, Barking, UK, 1981.

2.

ASTM D1895-96, Standard Test Methods for Apparent Density, Bulk Factor, and pourability of Plastics Materials, 2003.

.3. J. Dick and M. Gale in Handbook of Polymer Testing, Ed., R. Brown, Marcel Dekker, New York, NY, USA, Chapter 8. 4.

M. Boysen and J. Gronebaum, Plastverarbeiter, 1972, 25, 8, 549.

85

Mixing in Single Screw Extrusion

86

6

Solids Conveying in the Feed/Transport Zone

The three main stages of solids conveying, melting and pumping are equally important for overall throughput. On the one hand there can be no accumulations of material, and on the other, the output rate will be limited by the conveying efficiency of the worst performing zone. As problems will be cumulative with succeeding stages, the feed zone is all-important in terms of both overall output rate and output rate consistency. For many years, plastic extruder feed zones had smooth barrels, but feed zone inserts with grooves to restrain pellet rotation were introduced in Europe in the 1960s and have grown in application, particularly in combination with barrier screws, but less so in North America.

6.1 Smooth Feed Zones Forwarding by a screw (except in gravity conveyors) is bounded by two limiting cases. These are demonstrated in Figure 6.1. The bolt represents the extruder screw and the nut the bed of plastics granules in the screw channel.

Figure 6.1 Nut-on-bolt model for screw conveying. 87

Mixing in Single Screw Extrusion 1) If the bolt is turned and the nut is restrained from turning, the nut will move axially one pitch per revolution. i.e., we have 100% conveying. 2) If the bolt is turned and the nut is not restrained, it will turn with the bolt and not travel axially. i.e., we have barrel slip and zero conveying. A common comment on variable or reduced output rate is ‘there is a problem of screw slip’ whereas screw slip is essential for conveying and the observed variable or reduced conveying is due to ‘barrel slip’. The real forwarding situation is that particulate solids conveying is simulated in the model by very lightly restraining the nut so that it turns at, for example, half the turning rate of the bolt. It will then move axially at half a pitch for every bolt revolution. Two limiting cases can be applied to the screw channel in which a pellet is situated against the pushing face of the screw flight (Figure 6.2). 1) Pure axial movement

Where there is pure axial movement in which the pellets behave like a restrained nut on a rotating bolt, the pellet will be transported one pitch for every screw revolution in the axial direction V1. This is not possible in a smooth barrel.

2) Pure rotational movement

In the second limiting case, the material slips at the barrel surface and the pellet rotates with the screw in the direction V2 with no axial movement. In contrast to pure axial movement, this is possible in real situations.

Figure 6.2 Model illustrating possible conveying directions by considering a single pellet. 88

Solids Conveying in the Feed/Transport Zone The best possible situation that can theoretically be achieved is for pellet movement in a direction at right angles to the flight, V3. This will require zero friction between feed material and screw surface. In practice, the forwarding direction V4 will lie between V3 and V2 and therefore depend on the relative friction between both screw and barrel. High values of barrel friction providing drag, and low values of screw friction providing slip, will improve forwarding. This direction of forwarding, V4 can be defined by the transport angle as shown in Figure 6.2. The variables which determine forwarding efficiency and feed zone output in a given screw/barrel system are: 1) 2) 3) 4) 5) 6)

Pellet/barrel friction Pellet/screw friction Pellet bulk density Pellet/pellet friction Screw helix angle Screw channel depth

The most favourable conditions for solids conveying in a screw [1] are: 1) 2) 3) 4) 5)

High pellet/barrel friction Low pellet/screw friction High pellet bulk density Optimum helix angle (which is related to friction coefficient) A deep screw channel

When examining pellet behaviour, those that give the best hopper flow (e.g., die faced cut pellets that are almost spherical or suspension polymerisation beads) are more likely to give screw conveying problems. Irregular granulated scrap sheet or pipes which are more likely to bridge in the hopper will be conveyed more readily by the screw. This is because pellets that lock together easily will transmit forces applied by the turning screw flight to frictional drag forces at the barrel wall. This results in overall forward plug conveying. Conversely, beds of pellets that readily internally shear (or roll over each other), will not produce sufficient or consistent enough barrel wall drag to be easily conveyed forward. In general the coefficient of friction of polymer pellets against a steel surface will rise with increasing temperature, although the results in Table 6.1 show no significant rise up to 120 °C. The fall between 60 and 100 °C may be a ‘rubbing-in’ characteristic of the instrument. 89

Mixing in Single Screw Extrusion

Table 6.1 Polymer/Metal (External) Friction Coefficient at 0.5 rpm for a range of temperatures Test temperature Polymer

Form

60 °C

100 °C

120 °C

140 °C Initial

15 min

Coefficient of Friction HDPE

Pellets

0.086

0.083

0.076

0. 596

0.568

VHMWPE

Crumb

0.048

0.088

0.074

0.586

0.586

PP

Pellets

0.179

0.083

0.065

0.635

0.663

PP

Powder

0.112

0.086

0.081

0.865

0.681

6.2 Grooved Feed Zones In Section 6.1 the analogy of a nut on a turning bolt demonstrated that in the real situation the axial movement of the nut depended on the extent it was restrained from slipping between the fingers. This principle was introduced in the late 1960s and early 1970s by incorporating axial grooves in the feed zone to restrain pellet rotation [2, 3] (Figure 6.3).

Figure 6.3 Feed zone with reducing depth axial grooves. 90

Solids Conveying in the Feed/Transport Zone With the advent of very high molecular weight HDPE for applications such as large rigid blow-moulded containers, there was an impetus to improve the forwarding efficiency of the extruder feed zone as the feed material was a low coefficient of friction powder/crumb, with a bulk density less than other polyolefines. Polymer/polymer and polymer/metal coefficients of friction in Tables 6.1 and 6.2, although low, are overall very similar. Internal friction for very high molecular weight high-density polyethylene (VHMWHDPE) is comparable to HDPE and therefore the particulate bed should be equally resistant to internal shearing, a requirement for efficient grooved feed conveying. Friction data shown in Table 6.1 and Table 6.2 were produced using an ‘annular shear cell’. Bulk density values shown in Table 6.3 are significantly lower for VHMWHDPE except when comparing with PP pellets using containers having channel depths similar to laboratory extruder screw channels.

Table 6.2 Polymer/polymer (internal) friction coefficient at 7 rpm for a range of temperatures Test temperature Material

Form

Ambient

60 °C

85-90 °C

120-130 °C

Coefficient of friction HDPE

Pellets

0.49

0.46

–

–

VHMWPE

Crumb

0.50

0.48

0.48

0.50

PP

Powder

0.38

0.37

0.39

0.39

HDPE: High-density polyethylene VHMWPE: Very high molecular weight polyethylene PP: Polypropylene

Table 6.3 Bulk Densities for different particulates at three test container depths Depth of test container Test container

Form

80 mm

12.7 mm

6.3 mm

Bulk Density (g/l) HDPE

Pellets

510

485

450

PP

Pellets

518

480

393

PP

Powder

528

517

508

VHMWPE

Crumb

417

395

387

91

Mixing in Single Screw Extrusion Feed zones with axial grooves machined at regular intervals around the bore and extending 2-3 D (D represents the distance along an extruder screw in terms of screw diameters) beyond the feed zone opening will prevent the polymer particulates located within the grooves from turning. Internal friction between these particulates will transmit these forces through the bed of material, such that in an ideal situation, the solids material will be propelled like a restrained nut on a turning bolt. There is consequently a need for an adequate level of internal friction. The depth of the grooves becomes shallower towards the forwarding direction until they run out at the barrel surface. A grooved feed section sleeve is shown in Figure 6.4. A further refinement is to provide a taper to the feed zone bore over the same length as the grooves (Figure 6.5). The resulting compaction may increase internal friction which will further reduce any tendencies to internally shear. Very significant increases in output rate are evidently possible as shown in Figure 6.6, but very efficient water cooling is necessary to prevent frictional heat melting the pellets or crumb within the feed area. The results in this figure shows comparisons of output rates for PP pellets for a 38 mm extruder with a feed zone channel depth of 5.3 mm. Compared with the smooth bore feed, the grooved parallel bush increased output rate by 18% and the grooved tapered bush by 62%. Grooves are normally regarded as unsuitable for powders. Even so, the combination of grooves with tapered bush produced an increase in output for PP powder of 10% as shown in Figure 6.7 and Table 6.4. With PP pellets, the very high specific output rate achieved with the combination of granule feed and grooved feed resulted in effectively reduced usable output rates as a result of the conventional screw being unable to fully melt the polymer above a critical screw speed (Table 6.5).

Figure 6.4 Grooved feed zone sleeve used with 38 mm extruder. 92

Solids Conveying in the Feed/Transport Zone

Figure 6.5 Axially grooved, tapered bore, feed zone.

Figure 6.6 Extruder output rate versus screw speed: comparison of three feed zone sleeves using PP granules. 93

Mixing in Single Screw Extrusion

Figure 6.7 Extruder output rate versus screw speed: comparison of two feed zone sleeves using PP powder.

Table 6.4 Specific output rate comparisons for PP powder with PP pellets using three feed zone sleeves Specific output rate (g/screw revolution) Parallel/plain

Parallel + Grooves

Tapered + Grooves

PP Powder

3.9

3.6

4.3

Pellets

3.8

4.5

6.2

94

Solids Conveying in the Feed/Transport Zone

Table 6.5 Influence of grooves in the feed section on melting limit Feed Zone

Parallel/plain

Parallel + grooves

Tapered + grooves

Output rate (g/rev)

Melting limit (g/min)

Output rate (g/rev)

Melting limit (g/min)

Output rate (g/rev)

Melting limit (g/min)

PP Powder

-

-

3.0

206

3.6

250

PP Pellets

3.8

341

4.5

-

6.2

179

Overall output rates for VHMWHDPE were comparatively low: a specific output rate of 3.5 g/min being achievable only up to a screw speed of 50 rpm due to onset of surging also attributed to melting deficiencies of the general purpose screw. This was confirmed by model feed zone trials using an 8D screw in a short barrel with a range of applied back pressures which showed consistent output rates over a wide range of screw speeds when using a grooved feed. As feed zone channel depth is increased beyond a certain maximum, further increases in channel depth will result in increasing internal shearing of the solids bed until the forwarding rate reduces to a level similar to that for a smooth barrel surface [6]. Although the optimum depth with a grooved feed zone is significantly less than for a smooth feed zone, the forwarding is so efficient that a reduced depth and/or reduced pitch is used to limit the extra torque and feed zone cooling which is otherwise necessary. By using grooves in the feed zone, the output rate is driven and accurately controlled at the start of the screw and no longer dependent on the die resistance. The metering zone as a pump is therefore redundant and becomes available for part of the melting process and mixing the increased throughput [5]. It was therefore logical to use grooved feed zones in combination with barrier melting screws (see Chapter 7). However, the original consequence was that the extruder had to withstand very high feed zone pressures and drive torque. Power was dissipated by the intensive feed zone cooling required to prevent premature melting and breakdown of solids forwarding. There was also a requirement for increased wear resistant screws and barrels, particularly in the feed zone. Over a period of time, increasing polymer prices biased the objectives towards exploiting the pumping accuracy to achieve consistent extrusion dimensions. This enabled products such as pipes and films to be made at the minimum thickness allowed with minimised risk of falling outside the dimensional specifications. Consequently, raw materials’ costs were reduced. In more recent years the disadvantages of the need for intensive water cooling, wear, and high power consumption have been addressed. 95

Mixing in Single Screw Extrusion In more recent years, combining optimum channel depth with reduced screw pitch matched the feed zone’s specific output rate to the melting and mixing capacity of the remainder of the screw.

6.3 Particulate Friction Measurements With screw conveying efficiency being dependent on friction, both between pellets and steel, and within beds of pellets, friction data can be useful. Measurement techniques range from inclined planes to annular shear cells [6]. The friction data recorded in Tables 6.1 and 6.2 was obtained using a comparatively large instrument which was fitted within a large air circulating oven and driven by a vertical shaft from below. It was based on one at Warren Spring Laboratory which was in turn based on one devised for measurements on powdered coal. It consisted of a rotating annular trough containing the pellets under test into which was fitted either a smooth ring (for polymer pellet versus metal friction) or a ring with blades (for pellet versus pellet friction) The normal load was applied by adding weights to the ring (or ‘shoe’) and the transmitted frictional force measured by a load cell restraining the ring from rotating. The pellets were gripped by the very rough base of the trough and alignment of the shoe was ensured by the shoe’s location on the drive shaft. For more details see reference [7]. A smaller diameter instrument was later built which had a smooth heated shoe enabling higher normal loads to be applied. Pellets rubbed against a heated metal surface, steadily rising in temperature until melting started. An ‘expanded’ view of this instrument is shown in Figure 6.8. The larger instrument running under steady conditions was useful in comparing polyolefines in different particulate forms, particularly for grooved feed zone behaviour. Internal friction results were very consistent for specific polymers over the ambient to 120-130 °C temperature range, but varied from 0.38 for PP to around 0.45 for the HDPE/VHMWHDPE. External friction again showed differences between PP and HDPE/VHMWHDPE with no clear trends, except considerably greater variability with temperature. The sudden drop with the 60 °C to 100 °C tests for PP may be a ‘rubbing in’ effect. Figures 6.9, 6.10 and 6.11 show traces for temperatures and associated friction against time. The onset of wide amplitude friction traces indicate onset of melting of polymer in contact with the shoe. The examples show wide variations in friction characteristics with temperature between particulate forms. However, there are decreases rather than increases in friction up to the onset of melting. 96

Solids Conveying in the Feed/Transport Zone

Figure 6.8 Expanded view of an annular shear cell polymer to metal friction tester. (Drawing by Richard Humpidge. ©1975, Rapra Technology)

97

Mixing in Single Screw Extrusion

Figure 6.9 Friction changes with increased metal temperature for irregular HDPE granules. (©1975, Rapra Technology)

Figure 6.10 Friction changes with increased metal temperature for polypropylene powder. (©1975, Rapra Technology) 98

Solids Conveying in the Feed/Transport Zone

Figure 6.11 Friction changes with increased metal temperature for polypropylene powder preblended with 20% talc filler. (©1975, Rapra Technology)

6.4 Friction in the Feed Zone As low screw friction combined with high barrel friction should give the best solids forwarding, it appears logical that a combination of a highly polished screw with the matt finish of a cast iron feed section should be ideal. Other features such as feed zone screw cooling [8] might be of assistance. However, temperatures promoting faster melting, as found for a one piece barrel by Smith and co-workers [9] may be very effective. Friction measurements using a rising metal surface temperature as in Figure 6.9 show that coefficient of friction of pellets against steel are likely to steadily fall a little during early temperature rises and then rapidly rise as melting starts (Figure 6.9). These effects have also been shown by Huxtable and co-workers [6]. The use of polymer powder and the presence of particulate filler will influence this behaviour as shown in Figures 6.10 and 6.11. Feed zone screw cooling might be advisable if an additive causes pellets to stick to a warm screw surface. A similar example is that a quickly melting additive powder dusted on to polymer pellets may cause a breakdown in pellet conveying. A potential remedy is to get the extruder ‘up to speed’ on pellets alone to establish good solids conveying before the additive is introduced. Conveying factors also need considering before pumping liquid additives such as liquid colours, tackifiers and so on into feed/ solids conveying zones. 99

Mixing in Single Screw Extrusion The overall position is that feed zone problems need individual solutions, but that an understanding of the conveying mechanism is a step towards a satisfactory outcome.

References 1.

W.H. Darnell and E.A.J. Mol, SPE Journal, 1956, 12, 4, 20.

2.

G. Menges and R. Hegele, Plastverarbeiter, 1970, 21, 5, 11.

3.

U.M. Kosel, Plastics and Polymers, 1971, 39, 143, 319.

4.

R.W. Shales in Proceedings of a Rapra Technology Symposium on Screws for Polymer Processing: the Way to Better Productivity, Shawbury, Shrewsbury, UK, 1995, Paper No.8.

5.

J. Wortberg and R. Michels in Proceedings of the Annual SPE Conference – ANTEC ‘97, Toronto, Canada, 1997, p.48.

6.

J. Huxtable, F.N. Cogswell, and J.D. Wriggles, Plastics and Rubber, Processing and Applications, 1981, 1, 1, 87.

7.

J. Dick and M. Gale in Handbook of Polymer Testing, Ed., Roger Brown, Marcel Dekker, Inc., New York, NY, USA, p.171.

8.

E. Steward in Proceedings of the SPE Annual Conference - ANTEC, Dallas, TX, USA, 2001, p.62.

9.

W.S. Smith, R.A. Sickles, L.A. Miller and T.W. Womer, Proceedings of the SPE Annual Conference – ANTEC ‘07, Toronto, Canada, 2007, p.390.

100

7

Melting

7.1 Melting Mechanism As mixing cannot occur until the polymer has melted, delays in start of melting and the late completion of melting, are of concern with regard to thorough incorporation of additives. An extruder screw will contain unmelted pellets at the feed end and molten polymer at the discharge end prior to die entry, assuming the machine is functioning correctly. Between these two points there will be a transition from solid pellets to molten polymer over a number of screw turns, maybe occupying anywhere from one to two thirds or more of its total length. Melting has been studied in depth over the years and the technique usually used is that pioneered by Maddock [1] and by Street [2]. This technique had previously been used by Grant and Walker to study mixing [3]. The usual procedure is to run an extruder under steady conditions, stop the machine and apply rapid cooling to freeze the polymer in the screw channel. During this period the die and adaptor are removed. The screw is then ejected, usually with a hydraulic ram and sometimes with a short reheat to release the polymer from the barrel surface. Thompson and co-workers [4] used a clear tube to contain loose feed section material. Edwards and co-workers [5] used a barrel split axially to avoid risks of disturbing the surface during removal, whilst Edmondson and Fenner [6] used both methods. In order to visually inspect the transition from pellets to melt, mixtures of coloured pellets were used right from earliest investigations so that some appreciations of screw mixing behaviour was established concurrently with the melting studies. Following ejection, the exposed material filled screw can then be marked for further investigation by drawing a line along its length and numbering each screw turn. For ease of storing and sectioning it can then be unwound whilst just hot enough to be pliable and cooled as a flat straight length. The rate of melting can sometimes be gauged by examining the surface which may reveal the increasing width of the melt along the length of the strip. 101

Mixing in Single Screw Extrusion

Figure 7.1 Material spiral removed from extruder screw which also shows cut channel cross-section.

Viewing cross sections cut from the strip at regular intervals will show the transition from solid pellets to molten polymer. To show the mechanism more clearly, the extrusion is carried out using pellets comprising a blend of two colours or natural pellets dusted with pigment. Figure 7.1 shows part of a screw sample and the cross section. Unmelted pellets are delineated by a blue pigment, dry mixed before extrusion. The melting mechanism is typically as follows: 1) Screw conveying packs the pellets together and conveys them away from the hopper as a solid bed in which there is no relative movement between adjacent pellets (Figure 7.2(a)). 2) Granules of the solid bed pressed against and dragged along the hot barrel surface will melt and form a film (Figure 7.2(b)). 3) Providing the screw is not badly worn, i.e., there is a very narrow gap between flight and barrel, this molten film will quickly become too thick to pass over the flight and will be scraped off and forced inwards by the advancing flight (Figure 7.2(c)). 4) Moving down channel it will form a rotating melt pool at the back of the channel (Figure7.2(d)). 5) The remaining pellet solids bed will continue to advance down the channel such that the melting mechanism is maintained with continual addition to the melt pool. The rotating melt pool may also erode the solids bed at the interface. 102

Melting 6) We now have a situation in which the solid bed with no channel circulation coexists with a rotating melt pool. The solids may travel down channel faster than the melt pool, feeding it with material such that it becomes steadily wider [5] (Figure 7.2(e)). 7) Eventually all the granules will be melted and the rotating melt pool will fill the channel width (Figure7.2(f)). Thereafter the melt continues its helically flowing progress to the screw tip.

Figure 7.2 Diagram showing melting mechanism in screw channel.

7.2 Variations in Melting Rate Screw design is primarily aimed at achieving high and consistent output rate for a particular product from a particular polymer. A single flighted three zone screw as shown in Figure 7.2 has three main stages: 1) Feed/solids transport zone 2) Compression (channel depth reducing) melting zone 3) Metering or melt pumping zone. 103

Mixing in Single Screw Extrusion The channel depths and relative lengths of the three zones will depend on the polymer and possibly on the application. The compression ratio (or more accurately, the channel depth ratio) also varies depending on the polymer and application. Although it may appear logical that melting will start near the beginning of the compression zone and be completed at the end of this zone, the overall length and position will be influenced by temperatures, screw speed, and other variables. One important factor is the energy required to achieve melting. Figure 7.3(a) gives a few examples which show that in general semi-crystalline polymers such as high-density polyethylene (HDPE) require more energy for melting than an amorphous polymer such as polystyrene (PS). A diagrammatic representation of the two structures is shown in Figure 7.3(b).

(a)

(b)

Figure 7.3 Semi-crystalline and amorphous polymer structures. 104

Melting The melting/pumping zone generates the required pressure to push the melt at an adequate rate through the die. For example, HDPE has a comparatively low melt viscosity and requires a longer and/or shallower metering zone for pumping against a high back pressure die such as one used for small bore tubing or thin tapes. To avoid excessive shear heat development from a very shallow channel, a screw might be made with a deeper but longer metering zone at the expense of the feed and compression/melting zones. The overall effect is that the screw design is a compromise, where, in many cases melting with regard to mixing is given a low priority.

7.3 Solids Bed Break-up In addition to a melt film existing between solids bed and barrel surface, a melt film can also form between the solids bed and screw surface. This can promote a phenomenum called ‘solids bed break-up’. It was demonstrated by Edmondson and Fenner [6] where in addition to using both screw jacking and split barrel techniques, the melting rate was continuously monitored by observing pressure traces from a number of melt pressure transducers fitted along the length of the barrel. As the screw rotates, the pressure trace from a transducer on a continuous recorder will show a lower erratic pressure trace when pellets pass over the transducer diaphragm and a steady rise in pressure across the channel where molten polymer is present. Where both solids bed and melt coexist during melting, the two forms will appear on the recorded trace in proportion to the relative widths of pellet bed and melt (Figure 7.4). Melting progress can be plotted on a continuous basis by comparing the relative lengths of the respective pellet and melt length regions. The shape is influenced by the pressure transducer diaphragm diameter and response in relation to the screw flight width, pitch and pellet shape and size. Recorded traces including solids bed break-up were reproduced by Christiano and Slusarz in [7]. Edmondson and Fenner [6] found that acceleration of the solid bed occurred, causing it to break at regular intervals into separate pieces. These were one to four times their width and separated by melt regions which increased as the pieces moved downstream (Figure 7.5). This behaviour tended to occur towards the end of melting and could be prevented by applying screw cooling to freeze the melt film in contact with the screw. Screw cooling had been shown to improve mixing in earlier papers [1, 3] but this was a remedy used more often in the days when extruder screw length to diameter ratios (L/D) were typically 15:1 and the consequential reduced output rate due to the cooling was evidently acceptable. 105

Mixing in Single Screw Extrusion

Figure 7.4 Detection of solids and melt regions in a screw channel during extrusion.

Figure 7.5 Solids bed break-up (‘iceberg effect’). 106

Melting With a demand over the intervening years for higher output rates, the possibility of incomplete melting due to solids bed break-up needed attention. This lead to the development of melting devices and barrier screws.

7.4 Melting Devices A number of melting devices have been devised, having in common, a series of inlet and outlet channels separated by barrier flights such that all material must pass through the gap between the barrier flights and the hot barrel surface. Six of such devices have been reviewed by Rauwendaal [8]. They can be divided into those that have constant cross section channels, and those that have channels which taper either in depth or width or both. In all these cases they can be in-line causing a pressure loss, or angled to be pressure generating. It is likely that the inspiration for the first two (Street/Gregory of Egan [9] and Le Roy of Union Carbide [10] resulted from the early screw jacking research [1, 2]. Following a publication by Maddock [11], the Le Roy mixer [10] (commonly referred to as a Maddock mixer), has appeared regularly in publications on screw performance ever since and is often the standard by which other screw mixers are judged even though its principal role is that of a melting device. Although possibly not the best performing [8], its simplicity makes it possible for any toolroom to produce them easily and cheaply which presumably made it difficult for the patent holder to pursue infringements. In view of the sustained interest and applications of the Maddock element, (even being used in combination with modern barrier screws [12, 13]) some early work by the author examining its potential for black low-density polyethylene (LDPE) pipe extrusion using natural polymer plus carbon black masterbatch is described below [14, 15]. The element was used between flight turns 13 and 15 in a 24D, 38 mm extruder with what was otherwise a conventional screw with 8D metering and 8D feed zone sections. The Maddock element (shown in Figure 7.6) had four entry and four exit channels machined using a 9.6 mm ball end miller to give a channel width of 8.00 mm. The diameter was 38.02 and the alternate lands were ground to give a nominally 0.5 mm overall gap such that material entered an entry groove, passed over a barrier (through the gap) and left via the exit groove. All material must pass over the barrier to reach the die, but, unlike twin screw extruders, passes through a gap only once. With the addition of flightless entry and exit areas, the element occupied approximately the last 3D of the compression zone. Comparisons were made with a conventional three zone screw with the same feed and metering zone dimensions (Figure 7.2). Any semi-melted granules arriving at the element could not continue into the metering zone until melting was completed. 107

Mixing in Single Screw Extrusion

Figure 7.6 Maddock type barrier shearing element for 38 mm extruder. (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 7. ©1978, Rapra Technology)

Preliminary work showed that with the Maddock element screw, the melting/mixing efficiency and output rate of the extruder was very dependent on the temperature profile along the extruder barrel. Figure 7.7 shows comparisons of sections taken from the final turn of the extruder screw (no.24) for (a) increasing, (b) constant, and (c) decreasing temperature profile with LDPE. Reversing the barrel temperature profile so that the feed end of the screw was at the highest temperature produced the highest output rate and the best mixing.

Figure 7.7 Influence of screw temperature profiles on mixing with Maddock element at end of compression zone. Photomicrographs of screw tip channel section. (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 8. ©1978, Rapra Technology) 108

Melting

Table 7.1 Influence of barrel temperature profile on output rate with and without a Maddock element Mixing Temperature (°C) Conventional screw with screw Maddock Barrel Barrel Barrel Die 1 2 3 1 element

Screw speed (rpm)

Concentration Output of rate masterbatch (g/min) (%)

X

160

140

120

140

75

2.3

187

X

120

140

160

140

75

2.3

138

X

140

140

140

140

75

2.3

172

X

160

140

120

140

75

2.3

204

X

120

140

160

140

75

2.3

204

X

140

140

140

140

75

2.3

210

The results in Table 7.1 also show that whereas a reversal in temperature profile had a negligible effect on output rate of the conventional screw, the effect with the Maddock element was to increase output rate by about 35%. However, this was still about 8% less than for the conventional screw at the same screw speed. Photomicrographs of microtomed sections taken from an inlet and exit channel show that the element prevents unmelted or semi-melted material travelling any further (Figure 7.8). A lump or granule of masterbatch has been retained in the inlet groove where it would have subsequently melted. The combination of rolling and peeling should quickly melt and feed material over the barrier. However, when using a Maddock element at this point with polypropylene (PP), the unmelted polymer may exist as a result of it’s high enthalpy. This unmelted polymer obstructs flow across the barriers and severely reduces output rate. The unit’s mechanism was further illustrated by injecting a thread of pigmented LDPE from a 25 mm extruder through a pressure transducer tapping into unpigmented LDPE just before the element (Figure 7.9). Figure 7.10 is a photograph taken of material cut and opened out, and viewed from the inside face following screw freezing and ejection. The molten natural material entered the four inlet channels on the right of the figure, and after passing over the barrier, passed out through the four exit channels (on the left). The black ribbon was injected at the bottom left hand corner. As there was no screw flight in this area and the screw rotated past the stationary port, the ribbon was accepted alternately by each entry channel as it moved across. 109

Mixing in Single Screw Extrusion

Figure 7.8 Maddock element acting as barrier to passage of an unmelted masterbatch pellet. (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 9. ©1978, Rapra Technology)

Figure 7.9 Overall flow through Maddock element. (Reproduced with Permission from G.M. Gale, Masterbatch Flow patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 11. ©1978, Rapra Technology)

Figure 7.10 View from screw surface of unrolled material from Maddock element with injected black marker from side extruder. Marker injection point is also shown in Figure 7.9. (Reproduced with Permission from G.M. Gale, Masterbatch Flow patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 10. ©1978, Rapra Technology) 110

Melting

Figure 7.11 Rolling melts in inlet and outlet channels.

This resulted in the ribbon progressing as a series of separate ‘blobs’ progressing along each entry channel and passing over the barrier into the exit channel. In addition to a general overall forward flow, the material in the entry channel appears to have followed a rolling mechanism or spiral. The outer surface was peeled like a veneer and transferred over the barrier to the exit channel (as in Figures 7.8 and 7.11) where the combined forwarding and rolling continued. The transformation of the separate clear and black areas in the entry channels to totally black in the exit channels (when viewed from the outside) is the result of this rolling action producing a spiral lamina structure. Further trials using screw jacking following rapid barrel cooling compared sections from a conventional screw with the one fitted with the Maddock element. The feed material was LDPE with carbon black masterbatch. Figure 7.12 shows photomicrographs of microtomed channel cross sections taken every two turns and Figure 7.13 shows cross sections from the die adaptor. The overall effect is that in the early stages there was little difference between the two screws, but the better completion of the melting by the Maddock element is clearly shown. With the conventional screw, melting was incomplete two turns into the eight turn metering section, and cross sections from the die adaptor had poor overall homogeneity. There were indications of a substantial proportion of unpigmented natural polymer being fed to the die, particularly at higher screw speeds. Note: unmixed white area at 95 rpm. In comparison, the Maddock element ensured melting was completed, but laminar carbon black masterbatch streaks were still present although much thinner. There was also a reduction in specific output rate of about 8%. There has been a general trend to move the Maddock element out of the compression/ melting zone and re-position it in the metering section or sometimes at the screw tip. 111

Mixing in Single Screw Extrusion

Figure 7.12 Photomicrographs of screw channel cross sections comparing screws with and without a Maddock element (continued opposite). (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 13. ©1978, Rapra Technology)

112

Melting Figure 7.12 Continued …

113

Mixing in Single Screw Extrusion

Figure 7.13 Comparisons of strand die cross sections with and without a Maddock element for a range of screw speeds. (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 14. ©1978, Rapra Technology)

This restores the melting function to the compression zone with the Maddock element catching and melting solids bed break-up fragments and other material not quite fully melted. It also avoids blockage from unmelted polymer and consequent reduction in output rate. Several barrier elements have been devised having tapered inlet and exit channels such that their cross sections will be in proportion to the amount of material passing through at any particular point [8, 9]. A 38 mm element made to the drawing in Figure 7.14 increased loss in output rate by a further 7.5%, compared with the Maddock element and increased entry pressure from 7 to 9 MPa. Examination following screw jacking showed material flow was restricted to the central part of the barrier [14] (see Figure 7.14). 114

Melting

Figure 7.14 Material flow over barrier in element with tapered grooves. (Reproduced with Permission from G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978, Figure 18. ©1978, Rapra Technology)

7.5 Barrier Flight Melting Screws 7.5.1 The Barrier Screw Concept Research using conventional extruder screws has shown that: 1) Completion of melting of the solids bed moves towards and beyond the end of the compression/melting zone as screw speed is increased. It may even result in incomplete melting at the screw tip and lumpy extrusions. 2) The tip of the solids bed often breaks up so that unmelted clusters of granules travel downstream more quickly in the central channel regions. This is where shear strain mixing and melting conditions are at a minimum [6] (See Section 7.3). 3) The presence and size of masterbatch striations can increase with increasing screw speed as screw mixing capacity is reduced by later melting. Overall it has been clearly demonstrated that mixing cannot be completed until all the polymeric materials are fully melted and hence the later the completion of melting the less the opportunity for good mixing to be achieved. This applies whether the mixing is for masterbatched additives, polymer blending or uniformity of temperature of a single material. 115

Mixing in Single Screw Extrusion The barrier ‘melting screw’ can be an answer to this problem. They are understandably often referred to as barrier mixing screws because their replacement for conventional screws will often give an overall improvement in mixing as a consequence of their improved polymer melting. The barrier screw was patented in Europe in 1959 by Maillefer, a Swiss manufacturer of wire and cable machinery [16, 17] and curiously [8] patented by Uniroyal in 1961 [18]. A sketch of a Maillefer type barrier screw used by the author is shown in Figure 7.15. In this particular adaption for a 38 mm extruder it forms approximately the middle third of a 24:1 L/D screw with 8D feed zone and 8D metering zone. Cross sections of an unrolled channel in Figure 7.16 shows how a second flight with an increased pitch (the barrier flight) is divided off the original flight to form a second channel. The second channel steadily widens until eventually the barrier flight is re-united with the original flight. As a result, all material must pass over the barrier flight to continue. The gap between the barrier flight and the barrel wall is significantly greater than a normal screw clearance, allowing a molten polymer film to pass over, but is narrow enough to hold back unmelted and semi-melted polymer. Barrier clearances can be typically 0.5 mm [19, 20], whilst flight clearances are 0.1 to 0.2% of screw diameter.

Figure 7.15 Maillefer type barrier screw.

Figure 7.16 Representation of unrolled Maillefer screw channel. 116

Melting In Section 7.3, the solids melting process was described whereby during the final stages of melting, the remaining compacted solids could be enveloped by molten polymer causing break-up of the residual solids bed. These fragments became isolated from the hot barrel surface and travelled unmelted in the central regions of the channel with minimal mixing. By controlling the rate of melting, the barrier screw should both restrict melting to a designated region of the total screw and additionally avoid solids bed break-up problems.

7.5.2 Maillefer Barrier Screw The Maillefer barrier screw’s performance was first described in 1963 [21]. With the original situation that the barrier screw was patented in Europe by Maillefer, a Swiss cable machinery manufacturer, there was little interest in Europe in this technology other than by insulated wire and cable producers. When the patents lapsed, interest focused on designs emanating from developments which had taken place in the USA. In order to establish if better mixing of carbon black masterbatch into polyolefines could be achieved with this type of screw, comparisons were made by the author with a 38 mm laboratory extruder using the screw shown in Figure 7.15 and Figure 7.16, and with a conventional screw having the same geometry feed and metering zones. Figure 7.17 shows a 15.6% increase in output rate for the barrier screw over the conventional screw with LDPE and Figure 7.18 shows a decreased output rate of 32.7% with PP. Photomicrographs in Figure 7.19 of cross sections of strands extruded at 80 rpm show no significant improvement in mixing by using this particular barrier screw. The cavity transfer mixer is covered in Section 9.4. The reversal in comparative output rates between LDPE and PP for the two screws was due to the increased amounts of heat required to melt PP in comparison to LDPE. In Table 7.1 a number of polymers have been placed in order of their enthalpies with those requiring the largest amounts of heat to produce a melt at the top and the lowest at the bottom. The result is that semi-crystalline polymers will require the most heat and amorphous ones the least, and higher crystallinity more than lower crystallinity. Polyvinylchloride (PVC) is regarded as amorphous in this context. A high proportion of wiring is insulated with LDPE, plasticised PVC (PPVC), ethylene vinyl acetate and so on, which have low enthalpies and should benefit from this type of barrier screw. However, Maillefer cited the suitability of this type of screw for a wide range of polymers. These were PPVC, HDPE, LDPE, high molecular weight plyethylene (HMWPE), PP, polyamide-6 (PA6), PS, polymethylemethacrylate (PMMA) and cellulose acetate (CA). 117

Mixing in Single Screw Extrusion

Figure 7.17 Output rate versus screw speed comparing a Meillefer type barrier screw with a conventional screw for LDPE in a 38 mm extruder.

7.5.3 North American Barrier Screws With the barrier screw patented in the USA by Uniroyal, a rubber product manufacturing company, there was scope for development by screw supplying companies, although perceived infringements were vigorously pursued [8]. Compared with a conventional single channel screw, the introduction of a second channel introduced a large number of variables to choose from: 1) Barrier flight clearance. 2) Relative channel widths at any point. 3) Depth of solids channel at any point. 118

Melting

Figure 7.18 Output rate versus screw speed comparing a Meillefer type barrier screw wit a conventional screw for PP with a 38 mm extruder.

4) Depth of melt channel at any point. 5) Shape of transition from one channel to two channels: affects solid bed. 6) Shape of transition from two channels to one channel: affects melt and maybe residual unmelted solids. 7) Screw flight pitch in the melting zone, variations in pitch and changes in pitch of one flight relative to the other. 8) Overall proportion of screw length occupied by barrier flight, and starting and finishing points. 9) Provision of shearing and mixing elements after the barrier section. 119

Mixing in Single Screw Extrusion

Figure 7.19 Photomicrographs of extrusions containing 6% of a 40% carbon black masterbatch extruded at 80 rpm: Comparisons of four screw configurations.

Overall, the developments have been to depart from the single role of the barrier flight of separating melt from solids and hence controlling the melting model in what is otherwise a conventional screw. The parallel development of computer aided machining has enabled complex designs that would have previously been far too expensive to be economically viable. The separation of melt conveying from solids melting has enabled the channel sections at any point to be separately dimensioned to optimise their individual functions. Following the appearance of designs covering a diverse combination of the possible variables [22], screw designers appear to have focused on screws with a wider and shallower solids channel to increase melting rate combined with a narrower deeper melt channel to minimise shear heat development by the melted polymer. (Figures 7.20 and 7.21) The wider solids channel exposes the unmelted pellets to a larger area of hot melt film covered barrel surface against which the pellets are rubbed. This is a solids bed melting requirement previously described by Klein [23]. The varying channel depths can be easily combined with approximately constant channel widths for both solids and melt channels; the solids channel becoming shallower and the melt channel becoming deeper. This provides adequate forwarding 120

Melting

Figure 7.20 Barrier screw: modern configuration.

Figure 7.21 Modern barrier screw unrolled wit cross sections.

capacity and minimised shear heating. Overall, this leads to questions on how to start and finish the separate channels. The start of the second flight tends to cause a narrowing of the solids channel, which forms a restriction to solids conveying. This problem was highlighted by Hyun and co-workers [24] in screw removal experiments, and overcome by leaving a small gap at the start of the barrier flight which allowed a small quantity of solids to enter the melt channel. Similarly the closure of the solids channel may result in either blockage by unmelted pellets or, on the other hand, excessive shear heating. A similar solution to the channel entry restriction is to terminate the barrier flight without rejoining the main flight. In this case a rapid reduction in channel depth of the melt channel is necessary to bring it level with the metering zone channel depth. If the solids channel is too shallow near its termination, melt existing in the channel may be overheated and degraded by this high shear region; a problem which can occur in film co-extrusion when switching an extruder between different layers. The alternative appears to be to provide a margin of safety with a less shallow channel and provide a Maddock type barrier element in the metering zone to catch and fully melt any surviving semi-melted polymer. 121

Mixing in Single Screw Extrusion Apart from the barrier start and finish arrangements, the variables most likely to influence processing [20] are: Output rate: 1) Feed depth. 2) Solids channel end depth. 3) Melt channel depth. 4) Barrier flight clearance. Melt temperature: 1) Feed depth. 2) Barrier flight clearance. In trials by Steward and Braun [20], a 114.3 mm D, 24:1 L/D barrier screw was re-machined in stages to give the following: 1) Two feed zone depths. 2) Two solids channel end depths. 3) Three constant depth melt channel depths. 4) Two barrier flight clearances 0.51 mm and 1.27 mm. The polymers used were: 1) LDPE lower viscosity polyolefine. 2) Linear low-denisty polyethylene (LLDPE) higher viscosity polyolefine. The results can be summarised as shown in Table 7.2. It appears that the barrier screw concept originating over 40 years ago has reached maturity without becoming the screw of choice for many applications. Meanwhile, a range of quite different innovative screw designs have evolved in recent years to challenge both the conventional screw and barrier screw including the examples in Section 7.5.5. 122

Melting

Table 7.2 Overall influences of barrier screw variables on output rate and melt temperature Selected Variable

Output rate

Melt temperature

Increasing feed depth

Increased by 10-20%

Small reduction for LLDPE Negligible reduction for LDPE

Increasing solids end channel depth*

Small output increases

Negligible effect

Increasing flight clearance 5-10% increase for LDPE Decreased with LLDPE by about 15 °C at 75 rpm

Increasing melt channel depth

5% increase for LLDPE

Decreased with LDPE by about 5 °C at 100 rpm

Overall, gave 20% increase for LDPE

—

20-25% increase for LLDPE *Solids channel end width may have been narrower than some barrier screw designs, making the depth decrease less influential.

7.5.4 Combined Barrier Screws and Grooved Feed Zones The development of the grooved feed zone to improve forwarding of plastics pellets and crumb is described in Section 6.2. Originally, grooved feeds were used in combination with conventional screws with regard to melting, for low friction and crumb materials such as very high molecular weight polyethylene. Melting was achieved by a range of devices which achieved their required function with minimal influence on output rate due to the positive solids conveying. Eventually the barrier screw was substituted for the augmented conventional screw, and the combination of barrier screw and grooved feed zone became established in Europe whereas in North America barrier screws are mainly used with smooth feed zones [13]. The combination of barrier screws with grooved feed zones is a further variable which complicates comparisons of barrier screws with conventional screws. Although in 123

Mixing in Single Screw Extrusion general, grooved feed zones are far more likely to be used with barrier screws than with conventional screws. The overall development is summarised in Table 7.2. A comparison of a Maillefer type barrier screw with a conventional screw in Section 7.5.1, showed for the barrier screw, a reduced output rate when used with PP and an increased output rate when used with LDPE. Following the discovery that the metering zone of a Maillefer type screw was running only partially filled when extruding medium density polyethylene (MDPE), Qui and co-workers [25] evaluated a grooved feed zone to ensure the metering channel was filled. They confirmed that improved output rates could be achieved with LLDPE, PP and MDPE by using deep feed zone grooves, but that shallow grooves were unsatisfactory. Deep grooves were 9 mm wide x 4 mm deep at the beginning with a 63.3 mm D screw having a 9.9 mm deep feed zone channel. The shallow grooves were 1 × 1 mm.

7.5.5 Barrier Screw Developments The general form of the barrier screw has in general become established. They usually have a wide shallow solids melting channel and narrower relatively deep melt channel. The general needs of start and termination of barrier flight have also been established. Compression ratio is 2 to 2.5 [20], and the barrier gap (0.5 mm). Shear and pin elements are added as necessary. It has been generally established that the result will be higher and consistent output rates in most cases, combined with reduced melt temperatures. With the apparent approaching maturity of the Maillefer and Uniroyal derived barrier screw developments, further advances will most probably be in more radical innovations. Their potential diversity is illustrated by the following examples.

7.5.5.1 Barrier Screw with Divided Solids Melting Channel Observations by Christiano and Thompson [26] from screw push-outs showed that during the middle to late stages of the solids melting channel, the pellets were transformed from a bed of compacted pellets to a deformable homogenous compressed plug. A second flight was introduced to divide this later stage of the solids channel into two parallel channels with alternating periodic depth changes in a similar manner to the double wave screw described in Section 7.6.1. In this case it is applied only to unmelted but pliable material as molten polymer transfers over the main barrier flight into the melt channel as before. The undulating channels impart repeated compression and expansion transferring mechanical energy to the solids which speeds up the melting process. 124

Melting 7.5.5.2 Barrier Screw with Spiral Barrel Grooves for Overall Length [27, 28] This barrier screw appears similar to that in described [13]. With screw pitch decreased within the feed zone, air cooling is sufficient to control the feed zone temperature. Shearing elements plus a pin/pineapple mixer follows the barrier melting zone. The departure from the normal barrier melting arrangement is the provision of helical grooves in the barrel surface for its full length, but having a reducing depth in the feed zone. A separate grooved bush becomes unnecessary, but an axial grooved bush is still an option. The process works by utilising the barrel grooves to peel the melt film and softened pellets from the solids channel and transferring them into the melt channel. Additional advantages claimed include lower pressure demands on the feed zone such that wear as well as cooling needs are reduced.

7.6 Other Melting Screws 7.6.1 Double Wave Screw The Double Wave Screw has been around for many years following the patent by HPM in 1978 [29, 30]. The metering section consists of two equal width channels separated by an undercut barrier flight as in a barrier screw. However in this case the channels have varying depths along their length with continually repeating peaks and troughs which alternate between the two channels. When material approaches a peak from a valley, part of it will pass over the barrier flight into the adjoining channel. The slowing of any unmelted material by the peaks amid faster flowing melted polymer will promote their melting. Every half turn the process is reversed between channels which promotes distributive mixing. The overall output per revolution can be 30% greater than an equivalent conventional screw [31]. The general concept is shown in Figure 7.22.

Figure 7.22 Double wave screw: flow arrangement. 125

Mixing in Single Screw Extrusion

7.6.2 Barr Energy Transfer Screws The Barr energy transfer screw [32] has a barrier melting section following the feed section followed by the energy transfer section and a short metering section with a Maddock type shear element. The energy transfer section has two parallel channels but differs from the double wave screw by having alternating transfers across both flights by relieving the clearances for short distances. The variable barrier (VB) energy transfer screw [33] uses a generally similar mixing section to the energy transfer section but occupies about 50% more of the screw length and the barrier section is eliminated. The first part resembles a conventional screw with feed and compression zones, except that the metering zone is eliminated. However, the Maddock shear element is retained (see Figure 7.23).

Figure 7.23 Energy transfer and VB energy transfer screws. (Reproduced with permission from J.A. Myers and R.A. Barr in Proceedings of the 60th SPE Annual Conference – ANTEC 2002, San Francisco, CA, USA, 2002, Paper No.251. ©2002, SPE)

In blown film trials comparing a barrier screw and an energy transfer screw, and VB energy transfer screw, the latter outperformed the others in the extrusion of LLDPE and LLDPE blends with regards to both output rate and melt temperature. The extruder screw was 88.9 mm diameter, with 30:1 L/D.

7.6.3 Stratablend Mixing Screw This could have been included in Section 7.4 concerned with mixing elements which can be incorporated into many screw designs, but in trials using a 63.5 mm extruder by Somers and co-workers [34], the mixer length of 7D represented a third of the total length of the screw. Comparisons were made with a conventional screw using LDPE and acrylonitrile-butadiene-styrene in which black masterbatch pellets containing 30% pigment were premixed with white pellets containing 2% titanium dioxide in ratios of 35:1, 75:1, 100:1 and 220:1. 126

Melting

Figure 7.24 Schematic of ‘Stratablend’ mixing section, batch mixing for different screw configurations. (Reproduced with permission from S. Somers and M.A. Spanding, K.R. Hughes and J.D. Frankland in Proceedings of the Annual SPE Conference - Antec 1998, Atlanta, GA, USA, Volume 1, p.272. ©1998, SPE)

The screw mixing section consisted of a shallow screw channel into which three parallel rows of deep grooves extending for almost half a screw turn with small overlaps between each of the three adjacent rows and separated by lands formed by the base of the channel (see Figure 7.24). Cross sections of screw channel samples from ‘push-out’ trials showed significant improvements in both melting and mixing for the ‘Stratablend’ mixing screw in comparison to the conventional screw with relatively small increases in melt temperature. A slightly higher screw speed was necessary for the ‘Stratablend’ screw to match the output rate for the conventional screw, e.g., 80 rpm to match output at 74 rpm.

7.6.4 Shear-Ring Screw This screw by Wang [35], was designed to take advantage of the principle that reducing pellet size increases melting rate. The prototype described is basically a conventional screw with a series of interacting slotted rings located at intervals along the screw by pegs such that the rings can turn independently of the screw. The rings have slots on both inside and outside surfaces and can be angled with the same or opposite pitch to the screw or in-line with the screw axis. They can be fitted singly or in pairs.

7.7 Barrier Flight Screws versus Conventional Screws Considering that barrier flight screws date back to 1959, it is surprising that by 2001 general purpose screws remained the most commonly used design [36]. In spite of the barrier screw’s proven ability to produce significantly higher output rates at lower melt temperatures, the conventional screw is more often the screw 127

Mixing in Single Screw Extrusion of choice. The argument against barrier screws is that the design is specific to the polymer (usually polyolefin) and the application (frequently pipes and film). This is reinforced by the modern trend for production flexibility requiring screws to homogenise a wide variety of materials such that output rate and quality will not be met by a screw designed for a specific purpose. Bad news travels around the plastics industry a lot faster than good news and hence potential users will be far more aware of problems than successes. There is also the question of the higher costs of barrier screws, particularly when purchasing three or more extruders for a co-extrusion line. Examination of published information comparing four possible combinations of barrier and conventional screws with smooth and grooved feed zones gives some guidance although the varying selections of materials and equipment presents some difficulties in making comparisons. A summary of nine sets of results covering 18 polymers is shown in Table 7.3. In general the results show improvements in output rate when barrier screws are used - more so when combined with a grooved feed zone. Although polyolefins predominate, this is hardly surprising when considering both the large quantities extruded and the economic importance of high extrusion rates with commody polymers. With PA11, PA12 and styrene-acrylonitrile (SAN) for example, quality issues including those affected by downstream equipment may be more important. Even so, Maillefer’s reported output rates were considerably higher with the barrier screw for PA11 and PA12 than for the polyolefines used [21]. The maximum overall output rate with grooved feeds may also depend on the secondary reason that increasing pressure within the feed section may generate sufficient solids temperatures to cause pellet softening and a breakdown in conveying [37, 38]. These results show the ability of barrier screws to produce high output rates at lower melt temperatures, and when combined with grooves, to give linear output versus screw speed relationships. However, only one includes a comparison of a full range of pertinent mechanical properties. Panagopoulos [39] compared the machine direction tear, dart impact, puncture, and optical properties, including gels and haze, in LLDPE films made on a production blown film line using a barrier screw, comparing a smooth with a grooved feed zone. The results showed that in spite of the higher melt temperatures, the barrier screw in a smooth feed extruder had clear advantages in film quality and properties. With the grooved feed zone, pressure and therefore gauge variations were lower, and there was possibly less haze. It was concluded that higher stabilisation would reduce the effects of apparent overworking with the grooved feed zone, but at a high cost. 128

Melting As the grooved feed/barrier screw arrangement has been established for the extrusion of water and gas pipes where mechanical properties must meet well regulated pipe standards, the disadvantage discovered in the above film trials are not likely to be universal. However, the film trials emphasise the need to establish that physical properties necessary for the application are met, before accepting new technology in production plant. With so many possible design variables, there are inevitably inconsistencies between reported performances. Schoppner [36] showed that for PP the output rate depended entirely on whether or not a grooved feed zone was used. For the same feed zone the output rates for the conventional and barrier screws were the same. However, output rates from the grooved feed zone were 67% more than for the smooth feed zone. It also produced erratic changes in melt temperature with increasing screw speed for both conventional and barrier screws. Using only the smooth feed, the barrier screw increased specific output rate for SAN and PA by 10-15%. Only PP was used in both smooth and grooved feed zone extrusion trials [36]. The different combinations of screw and feed zone features mentioned in Table 7.3 are summarised in Table 7.4

129

Mixing in Single Screw Extrusion

Table 7.3 Comparisons made between melting/mixing devices with grooved and smooth feed zones for a range of polymers Screw Key

Polymer

Added Element

Smooth feed

Grooved Feed

Diameter (mm)

L/D

Shear

Pinned

LDPE

50

28:1







LLDPE

50

28:1







HDPE

50

28:1







PS

50

28:1







LLDPE

60/62

24:1

Mixing

Section

PVDF

150

28:1







PA11

135

27:1







PA12

135

27:1







LDPE

135

27:1







POP

63.5

24:1





LDPE

63.5

24:1





E [19]

LDPE

63.5

24:1



F [41]

ETPU

63

21:1



G [27, 28]

HDPE

50 75

34:1 36:1





HDPE

45

30:1

-



PP

45

30:1









SAN

45

30:1









PA6

45

30:1









FPVC

Maillefer

Screw



LDPE

Maillefer

Screw



HDPE

Maillefer

Screw



HMWPE

Maillefer

Screw



PP

Maillefer

Screw



PA6

Maillefer

Screw



PS

Maillefer

Screw



PMMA

Maillefer

Screw



CA

Maillefer

Screw



A [13]

B [39]

C [40]

D [7]

H [36]

I [21]

PVDF: polyvinylidine fluoride POP: poly(oxyethylene glycol) polymer Conv.: conventional

130

Barrier

Conventional



Barrier

Conventional











ETPU: engineering thermoplastic polyurethane FPVC: flexible polyvinylchloride

Melting

Table 7.4 Overall developments of screw and feed zone combinations Screw

Feed Zone

Last Zone

Conventional

Barrier

Smooth

Grooved

Metering

Melting/ mixing elements

1



-



-



-

2



-



-





3



-

-







4

-





-





5

-



-







It should be noted that: a) 1 and 2 are still widely used b) 4 is used more in North America c) 5 is used more in Europe d) With 3 and 5 (grooved feeds) there are examples where the metering zone is considered unnecessary for pumping/rate controlling through the die and some or all of it will be replaced with melting and/or mixing devices. e) Screw design is continually developing: for example, double wave and combinations, multibarriers and so on.

References 1.

B.H. Maddock, SPE Journal, 1959, 15, 5, 383.

2.

L.F. Street, International Plastics Engineering and Science, 1961, 1, 6, 289.

3.

D. Grant and W. Walker, British Plastics, 1951, 24, 8, 308.

4.

M.R. Thompson, G. Donoian and J.P. Christiano, Polymer Engineering and Science, 2000, 40, 9, 2014.

5.

M.F. Edwards, M.N. Gokhora and K.Y. Zayadine in Proceedings of the PRI Conference - Polymer Extrusion 2, London, UK, 1982, Paper No.17.

6.

J.R. Edmondson and R.T. Fenner, Polymer, 1975, 16, 1, 49. 131

Mixing in Single Screw Extrusion 7.

J.P. Christiano, and K.R. Slusarz in Proceedings of the TAPPI Polymers, Laminations and Coatings Conference, San Francisco, CA, USA, 1998, Book 1, p.483.

8.

C. Rauwendaal in Mixing in Polymer Processing, Ed., C. Rauwendaal, John Wiley, New York, NY, USA, Chapter 4.

9.

R.B. Gregory and L.F. Street, inventors; Frank W Egan and Company, assignee; US 3,411,179, 1968.

10. G. LeRoy, inventor; Union carbide, assignee; US 3,486,192, 1969. 11. B.H. Maddock, SPE Journal, 1967, 23, 7, 23. 12. J.A. Myers and R.A. Barr in Proceedings of the Annual SPE Conference – ANTEC, San Francisco, CA, USA, 2002, Paper No.251. 13. J. Wortberg and R. Michels, Proceedings of the Annual SPE Conference – ANTEC, Toronto, Canada, Volume 1, p.48. 14. G.M. Gale in Proceedings of the Annual SPE Conference – ANTEC, New Orlean , GA, USA, 1979, p.223. 15. G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978. 16. C. Maillefer, inventor; Maillefer, assignee; CH 363,149. 17. C. Maillefer, inventor; Maillefer, assignee; GB 964,428. 18. P. Geyer, inventor; Uniroyal, Inc., assignee; US 3,375,549, 1968. 19. T.W. Womer and G.L. Harrah in Proceedings of a TAPPI Conference Polymers, Laminations and Coatings, San Francisco, CA, USA, 1998, Book 1, p.475. 20. E. Steward and K. Braun in Proceedings of a Rapra Technology Seminar Screws for Polymer Processing: The Way to Better Productivity, Shawbury, Shrewsbury, UK, 1995, Paper No.7. 21. C. Maillefer, Modern Plastics, 1963, 40, 132. 22. C.I. Chung, Plastics Engineering, 1977, 33, 2, 34. 132

Melting 23. I. Klein, SPE Journal, 1972, 28, 8, 45. 24. K.S. Hyun, M.A. Spalding and J.R. Powers in Proceedings of the Annual SPE Technical Conference – Antec 1995, Boston, MA, USA, 1995, Volume 1, p.293. 25. D.Q. Qiu, P. Prentice and J.B. Hull in Proceedings of Rapra Technology Conference Screws for Polymer Processing II, 1998, Shawbury, Shrewsbury, UK, 1998 Paper No.5. 26. J.P. Christiano and M.R. Thompson in Proceedings of the SPE Annual Conference - Antec 2000, Orlando, FL, USA, Paper No.15. 27. E. Grünschloss, International Polymer Processing, 2002, 17, 4, 291. 28. E. Grünschloss in Proceedings of the Annual SPE Conference - Antec 2007, Cincinnati, OH, USA, 2007, p.405 29. G.A. Kruder and W.N. Calland in Proceedings of the SPE 48th Annual Conference – Plastics in the Environment: Yesterday, Today and Tomorrow Antec 1990, Dallas, TX, USA, 1990, Volume 1, p.74. 30. G.A. Kruder, inventor; HPM Corporation, assignee; US 4,173,417, 1979. 31. F.R. Pranckh, Molding Systems, 1998, 56, 3, 6. 32. J.A. Myers and R.A. Barr, in Proceedings of the 60th SPE Annual Conference – Antec 2002, San Francisco, CA, USA, 2002, Paper No.251. 33. J.A. Myers and R.A. Barr in Proceedings of the PLACE Conference and Global Hot Melt Symposium, Orlando, FL, USA, 2003, Paper No.17-3. 34. S. Somers, M.A. Spalding, K.R. Hughes and J.D. Frankland in Proceedings of the Annual SPE Conference - Antec 1998, Atlanta, GA, USA, Volume 1, p.272. 35. P.N. Wang in Proceedings of the Annual SPE Conference - Antec 2001, Dallas, TX, USA, Paper No.521. 36. V. Schoeppner, Kunststoffe, 2001, 91, 2, 32. [English translation, p.12] 37. G. Fuchs, Plastverarbeiter, 1968, 19, 765. 38. A. Schneider, Plastverarbeiter, 1968, 19, 797. 133

Mixing in Single Screw Extrusion 39. G. Panagopoulos, Jr., in Proceedings of the TAPPI Polymers, Laminations and Coatings Conference, Toronto, Ontario, Canada, 1999, p.703. 40. A. Wortberg in Proceedings of the 60th Annual SPE Conference – Antec 2002, San Francisco, CA, USA, Paper No.105. 41. T.A. Hogan, M.A. Spalding, K.S. Hyun, M.J. Hall in Proceedings of the Annual SPE Conference - Antec 1999, New York, NY, USA, 1999, p.124.

134

8

Screw Channel Mixing and the Application of Mixing Sections

8.1 Striations: Their Formation and Mixing in the Screw Channel Striation formation is an essential part of the distributive mixing process. In Chapter 2 it was shown diagrammatically how an individual pellet was transformed into a striation in a simple laminar shear field - the mechanism available for mixing in single screw extruders. It was also explained that the orientation in the direction of shear resulted in the decreasing effectiveness of the flow field in reducing striation thickness. The overall mixing requirement is to reduce the striation thickness to a value which represents an overall acceptable product whether visually or with respect to its physical properties. Clearly this is not always achievable: a wide range of striation problems were described in Chapter 1 and many examples have been described in published articles [1, 2]. The early investigations into melting behaviour and flow patterns in single screw extruders used mixtures of pre-coloured pellets. As these showed the presence of individually coloured laminar streaks, it was reasonable to assume that masterbatch striations were formed initially by the melting process and that once formed they could persist through the metering zone, adaptor and die, and into the extruded pipe, or cable, etc. With lateral stretching in blow moulding, striations in bottles and containers were particularly obvious. It can be predicted that every coloured pellet will be transformed into an individual striation, which will decrease in thickness and grow in length, but as in the model situation of Figures 2.12 and 2.13 (Chapter 2) the rate of extension (and decrease in thickness) will steadily decline as it journeys down the channel. It is also evident that pellets melting early have a greater opportunity for striation thinning than those melting later but with the decline in thinning as predicted in Figure 2.13, overall differences will be reduced. When higher output rates result in very late melting, there may be little opportunity for striation thinning to any degree. 135

Mixing in Single Screw Extrusion As the shear mixing occurs following melting, then in the route map of Chapter 4, we need to consider the mixing behaviour within the screw channel from the onset of melting through the remaining journey of melting completion and metering zone. Experiments carried out using a 38 mm extruder with a screw jacking facility verified that individual masterbatch pellets were transformed into individual finite striations during melting [3]. The extruder was run with the pellet feed maintained by hand at a level just covering the screw. Individual black masterbatch pellets were added at timed regular intervals. After a suitable overall extrusion time the extruder was stopped, rapidly cooled and the screw ejected with a hydraulic ram. A pipe grade low-density polyethylene (LDPE) was used with masterbatch pellets containing 40% carbon black. Cross channel sections confirmed that striations were formed during melting of individual masterbatch pellets. They showed that each pellet was transformed into a ribbon initially in the form of a tail which circulated around the channel cross section

Figure 8.1 Formation of laminar striation patterns by intervening plate and pipe die spider. 136

Screw Channel Mixing and the Application of Mixing Sections and also in a down channel direction. The ribbons formed a variety of patterns which were difficult to follow due to differing starting positions in the solids bed together with drag and pressure flow influences described in Section 8.3. As circulating laminar flow conditions continued for the full length of the screw, laminar streaks became longer and thinner, but persisted in the melt, and continued through the adaptor and both strand and tube dies. In so doing, patterns were formed depending on the intervening geometry (Figure 8.1: see also Figures 1.3 and 1.4).

8.2 Mixing During Melting Benkreira and co-workers [4] examined mixing during melting by measuring striation thickness of samples cut from the melting region of screw channel samples, removed following axial barrel splitting after cooling. The polymer was high-density polyethyelene (HDPE) and the feed was a 1:9 ratio of black pellets containing 5 wt% carbon black and natural polymer. Twenty µm thick microtomed sections were used with measurements of striation thickness at ×100 magnification using image analysis. Mixing assessment showed a high level of mixing on melting which then declined with little change thereafter except for the increase in size of the melt pool. Mixing during melting improved with increasing screw speed and to a lesser extent with decreasing barrel temperature, but this was no longer the case at the end of the screw. It is evident that the situation described in Chapter 2 showing how the striation thickness declines rapidly during initial shear strain but then reduces very slowly thereafter applies to the melting stages i.e., most of the mixing occurs during the very first stage of pellet melting.

8.3 Mixing in the Melt Filled Screw Channel Following completion of melting, the metering or pumping section of the screw pushes the molten polymer at a constant rate through the die. The screw channel is normally at its shallowest in order to provide the required discharge pressure although the depth and length have optimum dimensions to achieve the maximum output rate for a particular die head restriction and consequent back pressure [5]. This will be the cumulative pressure due to breaker plates, screen packs, feed pipes etc. This melt pumping stage is expected to provide polymer mixing whether for additive incorporation, blending or thermal homogenisation to give a uniform melt temperature. 137

Mixing in Single Screw Extrusion Tadmor and Gogos [6] have used a logical stepwise approach to modelling the mechanisms and relationships of screw pumping and mixing. They developed a screw in barrel arrangement in a series of steps starting from the simple moving infinite flat plate pump model. They then moved back a step to arrive at the established model of the unwound screw channel. The basic starting point is the laminar shear flow of a viscous fluid between a stationary and moving plate. There is a velocity gradient within the viscous polymer melt between the moving and stationary surfaces. The fluid is static at the surface of the stationary plate but moves with the surface of the moving plate. With the uniform velocity gradient associated with a viscous fluid between the moving and stationary plates, the material between moves in the same direction at a velocity dependent on its relative distance between the stationary and moving plates (Figure 2.7). The next stage is to confine the fluid between two walls, i.e., within a channel, and then impose a finite length with melt inlet at one end and a forming die at the other (Figure 8.2). As this arrangement for a pump will be restricted by the length of the plate, a continuous dragging moving surface is needed. This is achieved by curving the channel to form a semicircle fitted inside a rotating cylinder which replaces the flat moving plate (Figure 8.3). The design equation for this pump is:

Q=

V0 WH WH3 (P1 − P2 ) + L 12µ 2

Where Q = output rate

P1 = channel inlet pressure



P2 = channel outlet pressure



N = cylinder rotation speed



V0 = plate or cylinder velocity = πND



W = channel width



H = channel height



L = length



D = diameter

138

Screw Channel Mixing and the Application of Mixing Sections To make analysis easier, the equation is re-arranged to:

 V Q  P2 − P1 = 12µ L  0 2 −  3  2H WH  Where: P2–P1 is the pressure developed by the pump This shows: 1) Increasing channel width, W, increases pressure, but it is limited by inlet and exit problems. 2) Channel depth, H: at a given flow rate there is an optimum channel depth for a maximum pressure i.e., it depends on the extruder output rate and the die restriction: a drinking straw (high) or a water pipe (low). This optimum depth is given by:

H opt =

3Q WV0

3) Channel length, L: The longer the channel, the greater the pressure that can be developed and hence the higher the output rate that can be developed against the die restriction. With L limited by the sleeve diameter in the arrangement shown in Figure 8.3, the simple solution is to twist the channel so that it overlaps itself and winds into a spiral with a pitch of one channel width (plus wall) (Figure 8.4). This provides an ‘indefinite length’ without compromising the other variables. Fortuitously, the angled movement of the barrel surface with respect to the channel results in an extended flow path and hence shear strain which increases mixing. A rotating screw in a fixed barrel gives the same result as a fixed screw in a rotating barrel and is obviously the more practical arrangement for product extrusion. The feed inlet, die outlet, heaters, cooling fans or pipes etc., are fixed and a motor can rotate the screw. In considering the pumping/mixing relationship, the screw in the rotating barrel is now unrolled back to the straight channel and moving plate (as in Figure 8.5). The moving plate is no longer moving in the direction of the channel, but is skewed at an angle defined by the transformation from a ring to a spiral. If the spiral was the pitch of a typical extruder screw of one flight turn per screw length equal to one diameter, then the helix angle of the flight at the barrel surface will be 17.7 degrees (Figure 8.6). 139

Mixing in Single Screw Extrusion

Figure 8.2 Melt pumping a viscous fluid by dragging a plate over a channel. (Adapted from Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979)

Figure 8.3 Flat drag flow extruder rearranged into a semicircle to provide a continuous dragging surface. (Adapted from Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979) 140

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.4 Extension of channel length by rearranging to form a spiral. (Adapted from Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979)

Figure 8.5 Unrolled spiral showing barrel surface angled to channel and compared with unrolled half cylinder. (Adapted from Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979)

141

Mixing in Single Screw Extrusion

Figure 8.6 Flight helix angle.

Figure 8.7 Trajectory of a particle showing down channel flow. 142

Screw Channel Mixing and the Application of Mixing Sections As a result of the relative movement across the channel at an angle equal to the former helix angle for the spiral. movement of the plate at typically 17.7°, a tracer particle travelling down the channel near the plate surface will veer towards the flight. When reaching the flight at point A it will be forced downwards to the bottom of the channel at point B (Figure 8.7). It will then travel back across the channel near the bottom of the channel during it’s overall down channel journey to C and resurface at D when it reaches the opposite side. This spiralling path is then repeated. The viscous fluid movement between flights at the plate surface has to be balanced by a return journey at the bottom of the channel whilst moving down channel all the time. As the velocity gradient reduces from maximum at the moving plate/barrel surface (the drag flow) to zero at the stationary plate surface, the return path across the channel tends to be sensitive to outlet pressure resistance due to die restriction (pressure flow). This applies to both cross channel and down channel relative velocities. Figure 8.8 shows the cross channel circulating flow. At a point two-thirds the total depth, there will be no cross channel movement. Although shown as a closed loop, this is a section through a straightened melt flow spiral within the channel [6].

Figure 8.8 Cross channel flow behaviour.

The photomicrograph of a screw channel cross section in Figure 8.9 clearly shows the non-circulating layer at two-thirds of the channel depth. The overall trajectories will depend on the relative drag and pressure flow conditions as in Figure 8.10. Drag flow induced by the barrel surface moving along and slightly across the channel provides consistent down channel forwarding, but the returning flow in the bottom of the channel is sensitive to any restriction to flow, particularly die back pressure. With an open discharge a particle circulating near the barrel and screw surfaces will follow the trajectory shown in Figure 8.10(a) where there is only drag flow and no pressure flow. With a closed discharge, we get the particle trajectory shown in 143

Mixing in Single Screw Extrusion Figure 8.10(c) in which the particle circulates as shown, making no progress down the channel. At intermediate die pressures, the pressure flow will give trajectories intermediate between those for open and closed discharge (Figure 8.10(b)). This is the normal extrusion situation. It provides a combination of forwarding and mixing whereby mixing can be increased at the expense of forwarding by deliberately increasing back pressure. To achieve this, the barrel outlet is restricted with fine screens or a valve. This action is normally limited by both reduced productivity and increased melt temperature from shear heating. All viscous fluid mixing operations raise melt temperatures to a varying extent depending on the intensity and efficiency of the process. However, fine screens are often necessary for the removal of agglomerates, contaminants and hard gels, so any mixing improvement from the resulting increase in pressure will be a bonus.

8.4 Residence Time Distribution (RTD) If a tour group were to be shepherded by their leader from airport check-in to departure gate, they would arrive at the same time and would not be mixed up with other travellers. On the other hand, left to themselves, a few would get there quickly, others would take longer, and a number would tour the shops and arrive just in time to catch their flight. As a result the group would be well mixed with the other travellers. In the first situation there is a very narrow time distribution (or RTD) with little or no mixing, and in the second, a wide RTD and good mixing. In the extrusion situation the two extremes are achieved by: 1) Plug flow 2) Laminar flow as shown in Figure 8.11. Plug flow: if the walls of a channel were perfectly lubricated, we would get plug flow. All the viscous fluid would travel at the same rate with no shearing between layers. If an extruder could operate like this, there would be no mixing. Furthermore, if a second material was added for a fixed time, it would appear at the outlet over the same time period. There would be no increase in interfacial area, i.e., no mixing. Laminar flow: the velocity gradient causes an increase in interfacial area between adjacent layers, which is a measure of mixing which is taking place. If a second material was added at a fixed time it would appear at the outlet with a changing concentration over a period of time as shown in Figure 8.12. Data to produce such curves are easily obtained by practical experiments using tracers. 144

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.9 Photomicrographs of screw channel cross-section showing noncirculating layer at two-thirds depth. (Reproduced with permission from R.W. Shales, Mixing of Thermoplastics in Single Screw Rextruders, Department of Chemical Engineering, University of Bradford UK, 1989. [PhD thesis])

Figure 8.10 Influence of die back pressure on drag flow and pressure flow. (Adapted from Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979) 145

Mixing in Single Screw Extrusion

Figure 8.11 Plug flow and shear flow.

Figure 8.12 Residence time distribution (RTD).

A comparatively simple technique used by Bigg and Middleman was to introduce a dyed fluid into a nearly empty hopper and measure light transmission through the extrudate for samples taken at regular time intervals [7]. Wolf and White used radioactive tracers [8]. 146

Screw Channel Mixing and the Application of Mixing Sections

8.4.1 Concentration Smoothing The main advantage of RTD is that it smoothes out small variations in composition and consequently it is of particular interest in twin screw extrusion [9] where the polymer’s journey through the machine is in closed cavities for all or part of the barrel. The tightly intermeshing screws of counter-rotating twin screw extruders can form closed cavities for the length of the barrel, e.g., as used for single pass unplasticised polyvinylchloride powder compound extrusion for pipes, window profiles etc. The intermeshing screws of co-rotating twin screw compounding extruders will also form closed cavities, but to varying degrees overall depending on the configuration used. This applies particularly to the mixing cams, and use of reverse pitch etc. An additional factor in the case of single screw extruders is that the smaller the extruder, the further apart masterbatch pellets will be spaced in the feed zone in relation to the overall length of the screw channel [10].

8.4.2 Variation of Residence Time with Channel Position Referring to the mixing process mechanisms described in Sections 8.2 and 8.3, the variables of screw geometry and flow restriction effect down channel velocity profile but not those of the cross channel. Whilst the layers nearest the screw and barrel are spiralling around the top and bottom of the channel to give good lamina shear mixing, the area centred on the stationary (with respect to cross channel flow) two-thirds depth layer travels down channel at a faster rate. Furthermore a spiralling particle spends less of its time above the 2/3 h level than below it. The overall effect is that the 2/3 h region and a significant area on either side have a relatively short residence time, but the relative residence times near the top (barrel surface) and particularly in the bottom of the channel are much greater. However, the theory, (verified by Wolf and White [8]) shows that only about 5% of the output exceeds double the mean residence time.

8.4.3 Implications of Pressure/Drag Flow Effects The direct effect is that RTD smoothes out variations in additive concentrations, but that a long RTD (particularly as there can be a long tail as indicated in Figure 8.12), may result in thermal degradation, discolouration, gels, and excessive scrap when changing colours etc. 147

Mixing in Single Screw Extrusion The flow patterns which produce this effect in a single screw give a spread of laminar shear strain such that additives in central regions of the screw may have thicker striations than those that circulate around the regions nearer the screw and barrel surfaces. Elastic memory of the differences in shear history between material in the centre and outer screw channel regions may contribute to ripples in clear flexible polyvinylchloride sheet [11] and cause other rheological defects. The theory quoted assumed no leakage over the screw flights. Worn screws give better mixing, but at a disproportionate loss in output rate, illustrated in a graph by Colbert described by Lupton [12]. Both theory and practical experiments show that to be certain of adequate mixing of additives, conventional extruder screws need the addition of mixing sections to re-distribute, and (where necessary), re-orientate laminar striations.

8.5 Mixing Sections Having followed the polymer through the melting stage and melt pumping/metering stage we have a situation in which striations could well survive the limited mixing capability of the melt pumping stage, and pass on through the die into the product. In an attempt to rectify this situation, a very wide range of mixing sections has been devised to disrupt the ordered flow patterns within the screw channel. In addition to improving mixing, the designs have to take into account potential disadvantages of reduced output rate, excessive heat generation, polymer stagnation and potential cleaning problems. Erwin [13] commented, ‘These mixing sections are most remarkable for their variety. Vanes, pins, strange ducts of unusually shaped pieces of metal are incorporated in the melt channel to improve mixing.’ This causes a problem in choosing which mixer would be the best for a particular application. For example, a pineapple mixer has studs which may have some resemblance to a pineapple, but the studs may be square or diamond, large or small, be continuous like a chocolate bar or separated.

8.5.1 Maddock Mixer The Maddock element, described in Chapter 7.4, is often (and possibly inappropriately) used as the standard for judging other mixers. This mixer does contribute to distributive 148

Screw Channel Mixing and the Application of Mixing Sections mixing but it is primarily a melting device. An advantage for experimentation is that its geometry and design is comparatively firmly established. Even so the term ‘Maddock’ has been used to describe mixers resembling the original Le Roy patent in having inlet and outlet channels separated by a barrier but otherwise are completely different [14]. The Maddock type mixer shown in Figure 7.6 designed to fit a 38 mm extruder is typical of the many based on Maddock’s 1967 paper [15]. Experiments described in Chapter 7 showed that although mixing is improved by the Maddock mixer, the striations (which were divided into segments at the mixer entry) remained aligned for most of the time with the direction of shear.

8.5.2 Pins and Slots There are many possible arrangements of which two for pins are shown in Figures 8.13 and 8.14. To some extent, the same limitations apply to both pin and slot mixers and probably to most of the numerous ‘obstruction-flow-around’ mixers which have been devised. An understanding of the strengths and limitations of these mixers was comprehensively demonstrated by Martin. In a comparatively early paper [16], Martin reported findings from screw mixer samples (from screw jacking) and mixing performances for slotted flight mixers over a range of output rates. It was found that although mixing elements produced a new arrangement of streamlines by dividing striations, flow through or around interrupted reverse flights, was always lamina.

Figure 8.13 Mixing pins arranged axially. 149

Mixing in Single Screw Extrusion

Figure 8.14 Mixing pins arranged radially. (Reproduced with permission from D. Boes, Kunststoffe, 1974, 64, 11, 641. ©1974, Hanser Publishers)

Figure 8.15 Influence of mixing pins on polymer flow. (Reproduced with permission from D. Boes, Kunststoffe, 1974, 64, 11, 641. ©1974, Hanser Publishers)

150

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.16 Mixing quality versus output rate relationship for four screw arrangements. (Reproduced with permission from D. Boes, Kunststoffe, 1974, 64, 11, 641. ©1974, Hanser Publishers)

Similarly with pins, after the streams are split, they re-unite behind the pins so that. longitudinal mixing is only slightly improved compared to a screw devoid of a mixing section (Figure 8.15(a)) [16]. With slotted reversed flights and mixing plates, gaps need to be large enough to achieve both melt flow through the gaps and avoid excessive heat generation. Laminar shear mixing will be momentarily significantly increased due to the relative movement in the short channel formed by the gap between pins and the barrel surface (Figure 8.15(b)). In spite of these limitations, significant improvements were demonstrated (Figure 8.16), comparing two slotted reverse flight mixing screws with two standard screws of differing channel depths and a turbine type mixer as discussed in Chapter 9. Photomicrographs of sections covering a scale of 9 examples from good to poor mixing (1 = good, 9 = poor), were used for comparisons using a 30 mm 20D extruder over a range of output rates. The results show the shallower screw performed better than the 151

Mixing in Single Screw Extrusion deeper one, and the two reverse pitch slotted flight screws gave significant improvements in mixing. However, for these examples, mixing deteriorated considerably as output rate was increased. e.g., from rating 1 at 5 kg/h to rating 6 at 9 kg/h. This would have been due to the point of melting completion moving further downstream as speed increased. The same effect was found by Boes [1] for blow moulded parisons containing 0.8% masterbatch. Boes also found that the movement of melting completion towards the screw tip also caused a greater colour development in the outer zones of the parison.

8.5.3 Mixer Evaluation Using an Independent Drive When comparing mixing devices by measurement of striation thickness in the extrudate, under normal extruder arrangements the results are for overall mixing by both extruder screw and attached mixer. A particular variable is the influence of back pressure due to the mixer on both melting rate and mixing in the screw channel. A technique which avoids this problem is to decouple mixing from extrusion. An arrangement used by the author is shown in Figure 8.17 [17]. The primary objective was to obtain data on the influence of cavity size of the cavity transfer mixer (CTM) (described in Section 9.3). At the same time a variety of mixers were evaluated in an attempt to compare this particular mixer with a range of known mixing devices. An equal length of screw and a plain annulus (like the Couette model in Chapter 2) were used as controls. In addition to mixing performance, data was required on power consumption, temperature rise, and pressure drop.

Figure 8.17 Mixer evaluation arrangement using two extruder feeds. 152

Screw Channel Mixing and the Application of Mixing Sections Using this rig the mixer under test was rotated by an independent variable speed drive (with speed measurement), and fed by two 25 mm 20:1 L/D single screw extruders with identical screws running on polymer pigmented black and white respectively, both having been pre-compounded with a twin screw extruder. The torque was measured by strain gauges bonded to the mixer drive shaft with the signal recorded via a radio transmitter attached to the shaft. Mixer melt temperatures and pressures at the mixer entry and strand die were recorded, but there was no mixer cooling. There were two further advantages for this configuration: 1) Power consumption (from torque and speed measurement) can be more accurately measured by separating the mixer drive from that of the extruder. When measurements were made with a wattmeter attached to an extruder, the increase in power consumption due to the mixer was too small to give a significant measurement. 2) By varying the mixer speed independently of the extruder output, a better assessment of mixing performance can be made. Polymers used were LDPE, HDPE and polypropylene (PP). White material: Polymer compounded with 1% titanium dioxide. Black material: Polymer compounded with 5% of a 40% carbon black masterbatch. The mixers used (which fitted within a 32 mm barrel or stator) were as follows: Mixer A Figure 8.18. An annulus formed by a plain cylindrical barrel with a plain rotor (a Couette arrangement) Mixer B Figure 8.18. A short screw with pitch equal to diameter with a constant root diameter of 25.4 mm. Mixer C Figure 8.19. A cavity transfer mixer with 6 cavities circumferentially and 7 rows of cavities axially. Cavity spherical radius 7.5 mm. Rotor shown with one stator half. Mixer D Figure 8.19. A cavity transfer mixer with 3 cavities circumferentially and 5 rows of cavities axially. Cavity spherical radius 15.9 mm. Stator not shown. Mixer E Figure 8.20. A cylindrical rotor, root diameter 22 mm with 9 square pins circumferentially and 9 rows axially. The pins were nominally 5 mm square. 153

Mixing in Single Screw Extrusion Mixer F A similar mixer to the one described in reference [18]. Core diameter 25 mm with 12 blades circumferentially and 9 rows axially with adjacent rows displaced 15 degrees. Mixer G Figure 8.20. As mixer F had poor mixing performance, an axial gap of 3 mm was machined between adjacent rows of blades and its designation changed from F to G. Mixer H A matching rotor and stator with 5 axial semi-circular section channels in the rotor and 8 similar channels in the stator. Channel radius 4 mm. Mixer I

Figure 8.21. Similar to H with 8 channels in the rotor [19]. Stator not shown.

Mixer J Figure 8.22 (bottom). A pineapple mixer [20] with 15 rows of studs axially and 12 rows of studs circumferentially (Figure 8.25). Mixer K Figure 8.22 (top). A pineapple mixer with 7 rows of studs axially and 6 rows of studs circumferentially. Once mixing performance had been established, the overall performance became of less interest for the less commonly used mixers, and as a result, the generation of torque, pressure drop and temperature rise data for these mixers was not completed. The individual mixing performances were generally similar for the three polymers used. With constant extruder speeds of 50 rpm, the mixers were run at 0, 20, 40, 60 and 80 rpm. As comparisons between different mixers were very similar for the three polymers, results only for HDPE are shown. Figure 8.23 shows that the two cavity transfer mixers had the best mixing performance with either no visible striations or virtually none at 20 rpm. The mixing performances of the small and large cavity versions were very similar. Bar charts for temperature rise, pressure drop and power consumption are shown in Figures 8.24, 8.25 and 8.26. Drive power for annulus, screw, two cavity transfer mixers and a pin mixer were in the 80-110 watts range. Mixer G and pineapple mixers were 170 watts and above. Melt temperatures followed similar trends to drive power except for Mixer G. Pressure drop with HDPE was neutral for the screw at the selected speed, but was +0.69 MPa for LDPE and -0.345 MPa, i.e., pressure generating, for PP. 154

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.18 Mixer A (top), Mixer B (bottom). (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

Figure 8.19 Mixer C (top), Mixer D (bottom). (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

Figure 8.20 Mixer E (top), Mixer G (bottom). (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984) 155

Mixing in Single Screw Extrusion

Figure 8.21 Mixer I (Stator not shown). (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

Figure 8.22 Mixer K (top), Mixer J (bottom). (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

By noting from the photomicrographs such as those in Figure 8.23, the mixing speed required to eliminate striations, the mixing performances of the individual mixers can be compared. Comparisons of mixing performance using this criteria for HDPE, LDPE and PP are shown in Table 8.1 Comparisons of power requirements, heat generation and pressure drop were made from the instrumentation for each mixer. These values for mixing HDPE are compared in Table 8.2. The overall conclusions were that for the mixers in a smooth barrel (i.e., excluding C, D and I): 156

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.23 Photomicrographs of extruded mixer samples at varying speeds. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984) 157

Mixing in Single Screw Extrusion

Figure 8.24 Melt temperature rise across various mixers with HDPE. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

Figure 8.25 Drive power with HDPE. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984) 158

Screw Channel Mixing and the Application of Mixing Sections

Figure 8.26 Pressure drop across mixers with HDPE. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984)

Table 8.1 Comparison of mixing performances for three polymers Mixer

Mixer speed at which striations were eliminated (rpm) HDPE

PP

LDPE

A. Annulus (Couette)

›80

›60

›80

B. Screw

›80

›80

›80

C. CTM (small cavities)

20

20

20-40

D. CTM (large cavities)

20

20-40

20

E. Pins

40-60

60

60-80

G. 9 rows/12 vanes

60-80*

80*

80*

I. Fluted rotor/stator

40-80*

-

60*

J. Pineapple (small studs)

40

-

K Pineapple (large studs)

40->60

-

*These had a few striations which persisted over the indicated speed range

159

Mixing in Single Screw Extrusion

Table 8.2 Comparisons of heat generation, drive power and pressure drop for mixing HDPE Minimum speed (rpm)

Drive power (W)

Melt temperature rise (°C)

Pressure drop (MPa)

A. Annulus (Couette)

›80

90

-

-

B. Screw

›80

75

14

0†

C. CTM (small cavities)

20

105

22

1.035

D. CTM (large cavities)

20

95

17

0.69

E. Pins

40-60

85

16

2.415

G 9 rows/12 vanes

60-80

-

19

-

I. Fluted rotor/stator

40-80

-

-

-

J. Pineapple (small studs)

40

›180

33

5.52

K Pineapple (large studs)

40-›60

170

26

2.07

Mixer

†

As the screw was pressure generating, it ranged between positive neutral and negative for the three polymers.

1. The pin mixer was the best performing, with best mixing and overall lowest pressure drop, temperature rise and power consumption. It also proved very easy (and therefore of low cost) to make. The core was left as a polygon after turned flanges were slotted to form square pins. 2. The Pineapple mixer with the many small studs mixed much better than the one with fewer larger studs, but heat generation, power consumption and pressure drop (which are related), were very high. Possibly, variants resembling the pin mixer are better. Eitel and Funk [21] used a very similar arrangement to compare four mixers, again using a screw and a pin mixer. The mixers were as follows: 1) Standard screw, D 20 mm, channel depth 1 mm 2) Deep flighted screw with slots in the flights: D 20 mm, channel depth 4 mm 3) Shearing element 18 mm diameter 4) Pin mixer, diameter 20 mm, root diameter 16.4 mm 160

Screw Channel Mixing and the Application of Mixing Sections Direct comparisons were made for: Mixing speeds of 50, 80, and 100 rpm for the pinned mixer. Mixing speeds of 50, 80, and 120 rpm for the other three devices. Mixing element speeds 50, 80 and 120 rpm for 4 mixing elements. The material was LDPE, with polymer in one extruder coloured with 2% blue pigment. Mixing was assessed by sectioning extruded 10 mm rod and calculating standard deviation of striation thickness using image analysis. A graph of their results is shown in Figure 8.27. Standard deviation decreased (i.e., mixing improved) with speed and the pin mixer was again top of the mixing list as follows: 1) Pin mixer 2) Deep flighted slotted screw mixer 3) Standard screw 4) Shearing element However, the deep flighted screw improved considerably at 80 and 100 rpm to a similar value to the pin mixer. The steeper lines for the plain and slotted screws were attributed to their positive pressure build-up. Esseghir and co-workers [22] used a different arrangement of two extruders to isolate the potential varying influence of mixer back pressure on mixing, in which the mixer was attached to one of the extruders. A further difference was that incompatible blends of polystyrene and polyethylene were produced. These blends consist of droplets distributed within a continuous phase. In principle, for such a system, dispersive mixing is required to reduce the size of the droplets, whilst distributive mixing will normally occur concurrently with dispersive mixing. However, a shear strain (distributive) mixing route may occur as described in Section 14.2 Mixers evaluated were: 1) Maddock type with tapered inlet and outlet channels 2) Pineapple with diamond shaped studs 3) Twente mixing ring (TMR): a rotor and stator arrangement 161

Mixing in Single Screw Extrusion

Figure 8.27 Standard deviation of striation thickness against screw speed for 5 screw/mixer configurations. (Reproduced with permission from O. Eitel and R. Funk, Kunststoffe, 1991, 81, 1, 67. [English translation p.27]. ©1991, Carl Hanser Verlag)

In common with the first of these three investigations decoupling mixing from extrusion a rotor/stator type mixer (TMR) was included. This particular device is included in Chapter 10 (floating ring mixers). Overall, the TMR mixer produced the best results. The Pineapple mixer was better than the Maddock element. Table 8.3 is a comparison of results of mixer evaluations from eleven published papers. Simple comparitive ratings have been given where possible. However, with so many variables inevitively present, the complexity of making comparisons as indicated previously in this chapter means that a meaningful assessment requires close inspection of the original texts. 162

Screw Channel Mixing and the Application of Mixing Sections

Table 8.3 Summary of some published articles on mixing device performance Mixers

Polymer

Mix

Measurement

Maddock + long gridded element

HMWPE

Pigment masterbatch

Striations

Good results

[1]

ABS

Black and white pellets

Striations

Thinner striations obtained with screw + ET

[22]

Output rate + melt fracture

2 x output for ET

[23]

Screw Screw + ET Screw Screw + ET mixer Screw Slotted flight Shearing Pins Slotted flights and discs/pins Maddock Pins Pineapple Barrier Maddock Rhomboid 3D1D Rhomboid 6D1D Pineapple Screw Maddock Screw Pineapple Triple flight Z

LLDPE + LDPE

Polymer blend

Colour (two inputs)

Striations

1) Pinned mixer 2) Slotted deep flight screw

[20]

VHMWPE

0.2% masterbatch

Striations

1) Interacting pins 2) Pinned mixer

[24]

LDPE

Colour materbatch

Barrel window

All similar at screw tip

[25]

1) Pineapple 2) Maddock 3) Rhomboid 6C1D 4) Rhomboid 3D1D 5) Screw

[26]

1) Triple flight Z 2) Maddock 3) Pineapple 4) Screw

[27]

Striations

1) Reverse flight 5D (output = 3) 2) Reverse flight 2D (output = 4) 3) Screw (h = 1.5 mm, output = 2) 4) Screw (h = 2.2 mm, output = 1)

[15]

Striations

Striations thinner + better output compared with normal screw

[28]

HDPE

Yellow masterbatch

Striations

LDPE

Black masterbatch

Window in die and camera and image analysis

h = 1.5 mm LDPE

Slotted reverse flight 1D

Black masterbatch

Slotted reverse flight 5D

Strato- blend

Ref.

LDPE

Screw h = 2.2 mm

Rating*

ABS

Polymer blend

ABS: Acrylonitrile-butadiene-styrene ET: Energy transfer HMWPE: High molecular weight polyethylene LLDPE: Linear low-desnity polyethylene VHMWPE: Very high molecular weight polyethylene *Ratings listed and numbered in order of performance

163

Mixing in Single Screw Extrusion

References 1.

D. Boes, Kunststoffe, 1974, 64, 11, 641. [English translation p.9]

2.

R. Patsch, Kunststoffe, 1975, 65, 2, 89.

3.

G.M. Gale, Masterbatch Flow Patterns in Polyethylene Extrusion, Rapra Members Report No.16, Rapra Technology, Shawbury, Shrewsbury, UK, 1978.

4.

H. Benkreira, R.W. Shales and M.F. Edwards, International Polymer Processing, 1992, 7, 2, 126.

5.

E.G. Fisher, Extrusion of Plastics, 3rd edition, Newnes Butterworths, London, UK, 1976.

6.

Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979.

7.

D. Bigg and S. Middleman, Industrial and Engineering Chemistry Fundamentals, 1974, 13, 1, 66.

8.

D.H. White and D. Wolf in Proceedings of the 36th Annual SPE Conference – ANTEC, Washington, DC, USA, 1978, p.532.

9.

D.B. Todd, Plastics Compounding: Equipment and Processing, Hanser Publishers, Munich, Germany, 1998.

10. K. Luker in Proceedings of the TAPPI PLACE Conference, Boston, MA, USA, 2002, Session 31, Paper No.99. 11. G.M. Gale, Plastics News (Australia), 1983, No.12, 12. 12. A. Lupton, British Plastics and Rubber, 1997, February, 10. 13. L. Erwin, Polymer Engineering and Science, 1978, 18, 7, 572. 14. C-Y.A. Wong, T. Liu, J.C.M. Lam and F. Zhu, International Polymer Processing, 1999, 14, 1, 35. 15. B.H. Maddock, SPE Journal, 1959, 15, 5, 383. 16. G. Martin, Kunststofftechnik, 1972, 11, 12, 329. 17. M. Gale, Advances in Polymer Technology, 1997, 16, 4, 251. 164

Screw Channel Mixing and the Application of Mixing Sections 18. W.L. Krueger in Proceedings of the Annual SPE ANTEC Conference – Creating Value Through Innovation, Boston, MA, USA, 1981, p.679. 19. H. Teichmann, inventor; no assignee; US 2,810,159, 1957. 20. J.F.T. Pittman and G.L. Pitman in Proceedings of the PRI Conference Polymer Extrusion II, London, UK, 1982, Paper No.19. 21. O. Eitel and R. Funk, Kunststoffe, 1991, 81, 1, 67. [English translation p.27] 22. M. Esseghir, C.G. Gogas, Y. Dong-Woo, D.B.A. Todd and B. David, Advances in Polymer Technology, 1998, 17, 1, 1. 23. S.A. Somers, M.A. Spalding, J. Dooly and K.S. Hyun in Proceedings of the 60th Annual SPE Conference – ANTEC, San Francisco, CA, USA, 2002, Paper No.614. 24. J.A. Myers and R.A. Barr in Proceedings of the Annual SPE Conference – ANTEC, San Francisco, CA, USA, 2002, Paper No.251. 25. U.M. Kosel, Plastics and Polymers, 1971, 39, 143, 319. 26. C-Y.A. Wong and T. Lui in Proceedings of the Annual SPE Conference – ANTEC, Atlanta, GA, USA, 1998, Volume 1, p.284. 27. A.C. Rios and T.A. Osswald, M. del P. Noriega and O.A. Estrada in Proceedings of the Annual SPE Conference – ANTEC, Atlanta, GA, USA, 1998, Volume 1, p.262. 28. G. Harrah and T. Womer in Proceedings of the Annual SPE Conference – ANTEC, Atlanta, GA, USA, 1998, Volume 1, p.267. 29. S. Somers, M.A. Spalding, K.R. Hughes and J.D. Frankland in Proceedings of the Annual SPE Conference - Antec 1998, Atlanta, GA, USA, Volume 1, p.272.

165

Mixing in Single Screw Extrusion

166

9

Interacting Rotor/Stator Mixers

9.1 Overview A surprising number of patents exist for such devices, but only a few appear to have been used to any significant extent. It appears from the little information available, that in the majority of cases there is no clear disclosure on specific mixing objectives or achievements. In spite of this, it is known that distributive mixing has been achieved at a high enough level to justify the relatively high cost of these devices for demanding applications. The mixers known by the Author to have been used in production and in some cases still available are shown in Table 9.1.

Table 9.1 Interacting mixers used on production extruders Date

Inventor

Company

1955

T.A. Stanley

ICI

Rows of fixed and rotating teeth

[1]

1958

T.A. Stanley

ICI

Improvements to 1955 patent

[2]

1965

K.G. Gerber

Metal Box

Rows of overlapping key slots

[3]

1978

P. Renk

As Gerber with alternative construction

[4]

1980

G.M. Gale

Staggered overlapping hemispherical cavities

[5]

1987

B Schroter and co-workers

Straight rows of rectangular cavities with rounded ends

[6]

Barmag Rapra Reifenhauser

Design

Reference

167

Mixing in Single Screw Extrusion ICI and Rapra designs were made available to the extrusion industry via licensees. Barmag and Reifenhauser systems were made available to customers of their machinery. The Metal Box system was used ‘in-house’. A particular known attribute of the Barmag, Rapra, and Reifenhauser mixers is their ability to incorporate liquid additives such as polybutene film tackifier and liquid colours by direct injection. Their development is covered in historical order. The last three listed are logical developments of the first two, and are still being made after about 25 years since their inception. Two of the three suppliers are machinery manufacturers who have no reason to give away much technical information and hence for these two mixers, reliance is placed on two articles from Barmag and a collection of press release type articles from Reifenhauser. The author has therefore (but not entirely) relied mainly on his own work at Rapra Technology on the cavity transfer mixer (CTM). There are a few papers by others on this device but nothing from processors on any of these interacting mixers, but this situation is not unusual. A feature in common is their application to larger extruders. There appears to be no scale-up problems. Hensen’s table of sizes in 1984 [7] ranged from 30 to 200 mm diameter. By October 1986, Kobe Steel had produced (under licence from Rapra) four 305 mm CTM, part of a total of 29 ranging from 20 mm in steps of 10 mm, plus 11 at 65 mm [8]. In addition, pipe line CTM 350 mm and 400 mm diameter have been manufactured by Aspin Engineering [9]

9.2 Turbine Mixing Heads 9.2.1 Stanley (ICI) Mixer A turbine mixing head patented by ICI in 1955 [1, 2] enabled polyethylene pipe manufacturers to meet appropriate pipe standards when feeding pellet blends of natural polyethylene and carbon black masterbatch into the pipe extruder. Until then it was necessary to use the more expensive pre-compounded material. The turbine mixers were used by at least two large polyethylene pipe producers until being superseded by CTM, and at a later date, becoming unnecessary in some cases when price distortions and the polymer supply situation produced a switch back to extrusion of pre-compounded materials. 168

Interacting Rotor/Stator Mixers The turbine mixing heads consisted of alternate sets of fixed and moving teeth made by machining slots around the periphery of round discs: externally for the rotor and internally for the stator. The rotor formed a screw extension and the stator a matching barrel extension. The ICI 1955 patent [1] contains the simplest of sketches whilst the drawings (Figure 9.1) in the 1958 patent [2] covering improvements to the mixer in the 1955 patent, gives more detail but no construction details. The improvements included offsetting alternate sets of rotor or stator teeth to avoid pulsations, and angling the faces of the teeth to be pressure generating. The recommended space between the moving and stationary teeth was half the tooth thickness, but precautions were recommended to avoid engagement of fixed and moving teeth following thermal expansion. Good results had been experienced with a 50 mm mixer having 684 teeth and a 150 mm mixer having 948 teeth. Its potential application as an independently driven mixing device was included in patent claims.

Figure 9.1 Diagram from second ICI patent GB 787764 [1]. 169

Mixing in Single Screw Extrusion

9.2.2 Other Turbine Mixers In 1971, Kosel [10] described a turbine head (which appears very similar to the ICI mixer), and compared it with a pin mixer within a smooth barrel. Homogeneity and pigment distribution was judged by using a masterbatch at a very low level of 0.2% of the input particulate blend using virgin polyethylene in powder and crumb forms. Results showed the homogeneity of the extrusion with the turbine mixer over a wide range of screw speeds to easily outperform a pin mixer in a smooth barrel (Figure 9.2) A disadvantage was that output rate with the turbine mixer was unsatisfactory. It was planned to reduce the number of pins.

Figure 9.2 Mixing quality of difeerent mixing systems. (Reproduced with permission from U.M. Kosel, Plastics and Polymers, 1971, 39, 143, 319, Figure 22. ©1971, The Plastics and Rubber Institute)

During the following year, a paper by Martin [11] included comparisons of mixing using slotted flanges and pins on a screw both in a smooth barrel and a pinned barrel. With four rows of nominally square pins on both screw and barrel, and 12 pins in each row, distributive mixing deteriorated only slightly as output rate was increased. 170

Interacting Rotor/Stator Mixers With pins only on the screw, increasing output rate resulted in a steady decrease in mixing (Figure 8.16). This type of mixer has occasionally re-appeared including a comparison with a CTM when used with an independent drive [12] as described in the ICI patent. The turbine arrangement has also been combined with tapered roller bearings to prevent teeth engagement [13] but although mixing performance was good, this was accompanied by generation of unacceptable levels of shear heat [14]. The pin barrelled extruder on the other hand with pins located in the barrel wall passing through slots in the screw flight has been well established in the rubber industry.

9.3 Woodroffe Key Slot Mixers Although the turbine mixer produced the mixing required to meet polyethylene water pipe standards, problems could arise from two inherent features of the arrangement. 1) Differences in axial thermal expansion between heated barrel and screw could cause intermeshing of turning and stationary teeth, with consequential damage. Adjustment during heating from cold and start-up could result in long overall extrusion line setting up times. Hence the tapered roller bearing device referred to previously. 2) The layer by layer dismantling for cleaning and the subsequent re-assembly was time consuming. The Woodroffe key slot mixer avoided these problems.

9.3.1 Gerber (Metal Box) Mixer Gerber devised an add-on mixer intended to give the same mixing effects as the turbine mixer whilst avoiding the two disadvantages described previously [3]. The arrangement is shown in Figure 9.3 [3]. By having rows of slots similar to Woodroffe key slots arranged circumferentially both in the cylindrical rotor and similarly in a cylindrical stator with the rows displaced axially by one half pitch, the polymer alternated between cavities in the rotor and stator, with the cutting of the turbine teeth being achieved by the edges of the passing slots. The only published information is the 1965 patent [3], but these mixers (known as ‘Fred-heads’, after the engineer who made them) were used by the company on blow moulding machines. They gave good mixing of colour masterbatches and re-used bottle trimmings. 171

Mixing in Single Screw Extrusion

View of half stator

Figure 9.3 Metal Box key slot mixer. (Drawing from K. Gerber, US 3,174,185, 1965 [3]. Reproduced with permission from K. Gerber)

9.3.2 Renk (Barmag) Mixer Renk [4] used a similar geometry to Gerber but simplified the construction by machining slots into rings which were assembled on a rotor mandrel and in a stator sleeve to enable complete dismantling for cleaning. An arrangement is shown for the stator in Figure 9.4 [4]. Application of the mixer to fibre spinning has been described by Hensen [7] and by Dickmeiss [15]. 172

Interacting Rotor/Stator Mixers

Figure 9.4 Barmag Key Slot Mixer. (Reproduced with permission from F. Hensen, Advances in Polymer Technology, 1984, 4, 3, 339, Figure 15. ©1984, John Wiley and Sons)

The number of slotted cavities on both rotor and stator is 11 for the full range of screw sizes. The mixing effect was varied by the number of rings provided and the groove dimensions were varied to suit the material viscosity. The three basic arrangements from [7] and [15] shown in Figures 9.5 and 9.6 are as follows: 1) An extension to the extruder, i.e., a rotor attached to the extruder screw and a matching stator as a barrel extension. An additive injection port is shown at the mixer entry. 2) An independently driven mixer which has the flexibility to provide the required mixing for the injection of additive from a side feeding extruder. 3) An independently driven mixer with additive or dye injection and integral metering pump such that the mixing corresponds directly with the throughput rate. 173

Mixing in Single Screw Extrusion

Figure 9.5 Barmag options for feeding additives. (Reproduced with permission from F. Dickmeiss, Extrusion, 2007, 13, 2, 20, Figures 3 and 11. ©2007, VM Verlag GmbH)

Figure 9.6 Melt spinning diagram. (Reproduced with permission from F. Hensen, Advances in Polymer Technology, 1984, 4, 3, 339, Figure 20. ©1984, John Wiley and Sons) 174

Interacting Rotor/Stator Mixers The arrangements gave the flexibility for melt temperature control, either heating or cooling, and the rotor could be extended with a short length of screw to compensate for pressure drop across the mixing area. An arrangement for melt spinning is shown in Figure 9.6 in which a side extruder feeds masterbatch into a mixer fitted to the main extruder which in turn feeds a number of spinnerets. The gear pump metering the colour is coupled to the gear pump on the spinning heads in order to control concentration. In addition to achieving distribution of fine pigments and fillers, it was used for mixing antistatic agents, stabilisers and dyes. Polymers were polyamides (PA), polyesters and polypropylene. The photomicrograph in Figure 9.7 shows the very good distribution of 1% polyethylene oxide particles achieved in a 15 denier polyamide-6 carpet yarn. In 2007, over 100 of these units were reported to be in use for the production of fibres and films [15].

Figure 9.7 Photomicrographs of PA carpet fibres. (Reproduced with permission from F. Hensen, Advances in Polymer Technology, 1984, 4, 3, 339, Figure 21. ©1984, John Wiley and Sons) 175

Mixing in Single Screw Extrusion

9.4 Rounded Cavity Mixers 9.4.1 Rapra Cavity Transfer Mixer Although the term CTM is sometimes used to describe all mixers with overlapping stationary and moving slots and cavities, the term was initially used at Rapra Technology Ltd., to differentiate between the staggered row geometry of hemispherical cavities which followed the parallel row Woodroffe key slots used in the A2-B2 mixer described in Chapter 2. Unaware of the Gerber and Barmag patents, the slotted mixer described in Chapter 2 was developed as a successor to a combined roller bearing turbine mixer [13] which generated too much shear heat to be suitable for commercial polyethylene pipe extrusion. The results for this (A2-B2) mixer were sufficiently encouraging to carry out a trial on an industrial scale, replacing the toothed plates of an ICI type turbine mixing head fitted to a 120 mm extruder producing black low-density polyethylene (LDPE) water pipe to BS standards. The rotor and stator components were to be installed using the existing rotor shaft and stator housing. The objective was to achieve the level of mixing required whilst eliminating the problems of intermesh described in Section 9.3.

9.4.1.1 Cavity Transfer Mixer Scale-up From the outset, the cavities were comparatively large in comparison to key slots, as from an Erwin paper [16] the cutting and turning mechanism was an adjunct to the lamina shear mixing for the end result of good distributive mixing. When scaling up from a 38 mm laboratory extruder to a 120 mm production extruder producing mains water pipe to British Standards there was concern that the following points needed attention: 1) Restriction to polymer flow should be minimal. 2) Lands between cavities should be minimised to avoid excessive generation of shear heat. 3) Minimise risks of stagnation. 4) Minimise heat generation resulting from increased shear rates generated by larger screw diameters of production extruders. Higher extrudate melt temperatures can slow down production by exceeding cooling capacity of downstream equipment. 176

Interacting Rotor/Stator Mixers 9.4.1.1.1 Shear Heating The designs of the Metal Box and Barmag mixers appear to emphasise the cutting action, whereas theoretically there is an optimum balance between shear mixing and cutting/turning stages [16]. The shear rate in a screw channel is given by:

S= where

πDN h

S = shear N = screw speed D = barrel diameter h = channel depth

As high shear rates produce heat and hence normally undesirable rises in melt temperature, h the channel depth should be as deep as possible. This will also apply to mixer cavities.

9.4.1.1.2 Pressure Drop Pressure drop across a slot or rectangular channel is given by:

∆P = K1

QhL Wh3

And for a circular hole by:

∆P = K 2

QhL πR 4

Where ∆P = pressure drop K1 and K2 = constants Q = volumetric flow rate h = viscosity L = length of slot or hole W = width of slot R = radius of hole h = height of slot This assumes R and h are small compared with L and W. 177

Mixing in Single Screw Extrusion This means that the smallest cross section of the cavities if they are approximately rectangular (as in the case of key slots) or the radius (as in the case of half cylinders or half spheres) will have a disproportionate effect on the restriction to flow, Therefore, such dimensions should be as large as possible consistent with minimal risks of stagnation. With the hemispherical cavities described below, the overall interconnecting path can be roughly approximated to a series of round tubes and hence a small increase in radius will greatly reduce the restriction to flow. Overall there is a need, consistent with good mixing, that the shear rate is as low as possible, the pressure drop is as low as possible and the streamlining to be very good (See comparisons of mixers C and D in Section 8.6). All these features must also take into account the need to minimise risks of polymer stagnation in deep cavities and fit the openings close together in both internal and external surfaces. In scaling up from 38 mm to 120 mm, the cavity arrangement was considered as follows: 1) When cavities are opposite cavities, shear rates will be low and pressure drops will be low. 2) When cavities are opposite lands, shear rates will be higher and pressure drops disproportionately higher. 3) The greater the cavity overlap between opposite rows, the lower the pressure drop. 4) Rotor cavities have to be fairly well spaced circumferentially to provide adequate wall strength. The deeper the cavity the wider the rotor lands must be. 5) Deep cavities minimise pressure drop and shear rate but minimise laminar mixing and are more likely to produce stagnation and cleaning difficulties. 6) Shallow cavities give good mixing with lower risks of stagnation but at the expense of high shear rates and pressure drops. 7) Rounding of corners to improve streamlining increases the land area and potential heat generation. Note that theoretical treatments of this mixer configuration have been made by Bromilow and Hulme [17], de Jong [18] and Wang and Manas Zloczower [19]. The conflicting features of changing cavity shape to minimise land area, maximise cavity volume, avoid stagnation from corners and depth are shown in Figure 9.8. 178

Interacting Rotor/Stator Mixers

Figure 9.8 Compromises with scaling-up overlapping parallel cavity rows with key slots. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Technology Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 3.4. ©1984, Rapra Technology)

It was decided that the CTM should be nominally the same diameter as the extruder barrel (but minimally larger) such that the extruder screw would pass through the stator should this removal procedure be required whilst larger diameters might exceed torsional strength of the screw connection. If, in scaling up, the screw speed is unchanged, the output rate of an extruder will rise by the cubed power of the screw diameter. Based on this assumption, cavities consisting of split spheres would be appropriate. In practice screw speed is reduced as 179

Mixing in Single Screw Extrusion diameter is increased and hence the size of the spheres has been scaled up according to screw channel depth scale-up rules. The use of hemispherical cavities enables cavities to be used in a staggered configuration as shown in Figure 9.9. The potential advantages were as follows: 1) At any point during rotation each cavity is open to three opposite cavities with a large overlapping area being possible. This minimises pressure drop, and variation in cross sectional areas during rotation has a small amplitude at a high frequency. 2) Land lengths are minimal so that shear rates are minimised, i.e., total channel depth is the sum of the opposite cavity depths (Figure 9.10). 3) The configuration is streamlined with no corners for stagnation to occur. 4) Shape is conducive to circulatory flow which can link with opposite cavities. 5) Within geometric constraints, a wide range of cavity sizes can be used. 6) Machining is comparatively easy. 7) Cleaning and polishing can be carried out easily with rotary tools.

Figure 9.9 Overlapping hemi-spherical cavities. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Technology Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 3.5. ©1984, Rapra Technology) 180

Interacting Rotor/Stator Mixers

Figure 9.10 Streamline flow through and within overlapping cavities. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Technology Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 3.6. ©1984, Rapra Technology)

Figure 9.11 38 mm mixer used to evaluate performance of overlapping circular cavities. (Reproduced with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Technology Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figures 3.7 and 3.8. ©1984, Rapra Technology) 181

Mixing in Single Screw Extrusion To evaluate the proposed geometry, a 38 mm mixer was constructed as shown in Figure 9.11. For ease of construction, the same housing was used for the stator as used for the A2-B2 mixer described in Chapter 2. The stator consisted of a machined cylinder having holes bored in the required pattern to form the circular open cavities which were blanked off to form closed cavities by a sliding fit within the cylindrical mixer body. The rotor was a solid extension to the screw with cavities machined using a ball-end milling cutter. Figure 9.11b, (in which the rotor has been inserted into the stator), shows how the cavities overlap. 9.4.1.2 Assessment of Mixing Performance by Striation Thickness Measurement A repeat was made of the experiment described in Section 2.3.4.2, in which clear flexible polyvinylchloride (PVC) was extruded with a 38 mm extruder and a striation of flexible black PVC was injected from a 25 mm extruder through a transducer port just before the mixer entry. Following stabilisation of extrusion conditions the extruder was stopped, the 8 mm strand die removed and the rotor and stator removed

Figure 9.12 Striation thickness ratio versus mixer stage for A2-B2 and round cavity mixers. (Adapted with permission from G.M. Gale, Development of the Cavity Transfer Mixer for Plastics Extrusion, Rapra Technology Members Report No.104, Rapra Technology, Shawbury, Shrewsbury, UK, 1984, Figure 4.1. ©1984, Rapra Technology) 182

Interacting Rotor/Stator Mixers together by ejecting the screw with a hydraulic ram. Following separation from the rotor, the PVC was easily removed from the stator cavities as they were open on both sides. Microtomed sections were prepared and striation thickness measured using a microscope fitted with a graticule. A graph was drawn of to/tc against S where: to = striation thickness at mixer entry tc = striation thickness in the cavity S = number of cavity row (representing distance along the mixer) Using a log versus linear scale, the resulting graph had a straight line relationship as before but with a steeper slope (Figure 9.12), i.e., mixing performance was even better. 9.4.1.3 Production Trials With a satisfactory performance from the 38 mm CTM, a 120 mm unit was fitted as a replacement for the toothed plates of an existing turbine mixer attached to the production extruder producing LDPE and medium-density polyethylene pipes to British Standards (Figure 9.13). The unit ran satisfactorily in all respects and very significantly reduced

Figure 9.13 Rotor and stator halves used to replace toothed discs in 120 mm turbine mixer extruding PE pipes. (Photograph taken by G. Lawley. ©1980, Rapra Technology) 183

Mixing in Single Screw Extrusion

Figure 9.14 CTM (45 mm) for extruding XLPE pipes. (Photograph taken by G. Lawley. ©1990, Rapra Technology)

start-up times compared with using the turbine mixer. A photomicrograph is shown in Figure 9.14. This was followed by a production trial using a 114 mm vented compounding extruder where it enabled it to produce black polypropylene injection moulding compounds and other compounds to meet a range of standards. This prototype CTM was used for 24/7 production for 11 years before being scrapped with the extruder. Following further laboratory investigations, the cavity sizes and clearances were increased for future mixers and the number of rows standardised at 4 or 5 for most applications (Figure 9.14). An unusually long CTM with 13 cavity rows for a special application is shown in Figure 9.15.

9.4.2 Reifenhauser Staromix The Reifenhauser Staromix has rectangular cavities with rounded ends, arranged in parallel rows, the rotor and stator rows being a half pitch apart to provide a continuous path similar to the Gerber and Barmag mixers ([6] and Figure 9.16). 184

Interacting Rotor/Stator Mixers

Figure 9.15 13 row production CTM for special application. (Photograph taken by G. Lawley. ©Rapra Technology)

Figure 9.16 Reifenhauser Staromix. (Reproduced with permission from Reifenhauser ©1988, Reifenhauser)

185

Mixing in Single Screw Extrusion

Figure 9.16 (continued) Reifenhauser Staromix (Drawing based on patent [9.6])

In [20, 21] the Staromix was combined with a two stage screw utilising a flight pitch change at the end of a grooved feed/transport zone and also included a peg mixer. This enabled additive introduction at the end of the barrel which reduced wear and increased the range of both additives and raw materials which could be used. Concurrently, film down gauging could be increased whilst retaining properties. This screw/mixer combination could also be used for incorporating polybutene tackifier into blown and cast film, and achieve good masterbatch mixing during pipe extrusion.

References 1.

T.A. Stanley, inventor; ICI, assignee; GB 787,7649, 1955.

2.

T.A. Stanley, inventor; ICI, assignee; GB 841,743, 1958.

3.

K.G. Gerber inventor; Metal Box, assignee; US 3,174,185, 1965.

4.

P. Renk inventor; Barmag, assignee; US 4,253,771, 1978.

5.

G.M. Gale, inventor; Rapra lib, assignee; US 4,419,014, 1983.

6.

B. Shroter, W. Schwarz, R. Fellman and H-J. Bartels, inventors; Reifenhauser, assignee; US 4,913,556, 1989.

7.

F. Hensen, Advances in Polymer Technology, 1984, 4, 3, 339.

8.

K. Inoue and S. Fukumizu in Proceedings of the Rapra Technology Symposium - Making the Most of the Cavity Transfer Mixer, Rapra Technology, Shawbury, Shrewsbury, UK, 1985, Paper 9.

9.

A. Aspin, Aspin Engineering Ltd., (Private Communication), 2008.

186

Interacting Rotor/Stator Mixers 10. U.M. Kosel, Plastics and Polymers, 1971, 39, 143, 319. 11. G. Martin, Kunststofftechnik, 1972, 11, 12, 329. 12. W.W. Mumford in Proceedings of the Rapra Technology Symposium Making the Most of the Cavity Transfer Mixer, Rapra Technology, Shawbury, Shrewsbury, UK, 1985, Paper 4. 13. G.M. Gale, inventor; Rapra, assignee; GB 2,048,701, 1979. 14. M. Penny, Investigation of a Roller Bearing Mixer for Extrusion Compounding of Carbon Black with Polyolefins, Rapra Members Report 45, Rapra Technology, Shawbury, Shrewsbury, UK, 1980. 15. F. Dickmeiss, Extrusion, 2007, 13, 2, 20. 16. L. Erwin, Polymer Engineering and Science, 1978, 18, 7, 572. 17. T.M. Bromilow and A.T. Hulme in Proceedings of the Rapra Technology Symposium - Making the Most of the Cavity Transfer Mixer, Rapra Technology, Shawbury, Shrewsbury, UK, 1985, Paper 3. 18. E. de Jong in Proceedings of the Rapra Technology Symposium - Making More of the Cavity Transfer Mixer, Rapra Technology, Shawbury, Shrewsbury, UK, 1988, p.14. 19. C. Wang and I. Manas Zloczower, International Polymer Processing, 1996, 11, 2, 115. 20. J. Gale, Packaging Week, 1991, 6, 33, 14. 21. Reifenhauser News, 1988, December, p.13

187

Mixing in Single Screw Extrusion

188

10

Floating Ring Mixing Devices

10.1 Introduction Floating ring mixers for extrusion use an interactive rotor and stator in which the stator floats as an uncoupled internal sleeve within the bore of the barrel. They are derived from the mixing version of the check ring used at the screw tip in an injection moulding machine. The potential disadvantage of reduced mixing capacity due to limited stator bore and short length can be more than compensated by the high mixing performance of interacting moving and fixed elements. In the extrusion version, retrofitting will require substitution of the last few screw turns, thereby reducing the effective screw length.

10.2 Injection Moulding Check-ring Mixers It is not surprising that the melting and mixing limitations in single screw extruders can also occur in injection moulding, as the screws used are very similar and often shorter than those used in extrusion. If the mould volume is large in relation to the maximum shot capacity, then the length of screw available for both melting and mixing will be short when fully retracted. On the other hand, as the cooling time is usually longer than pre-plasticisation time, the back pressure can be increased to aid mixing, provided the cooling time is not increased as a result of additional heat generation by the screw. Maddock and pin mixing elements can be used, but the more efficient interacting pegs or cavities cannot be used in the same manner as for extrusion described in Chapter 9. The rotor and stator would move completely out of register as the screw reciprocates. Although techniques have been devised, they required greater space than is normally available [1, 2]. During the injection part of the cycle, backflow up the screw channel by the molten polymer is prevented by some form of non-return valve such as a ‘check ring’ (Figure 10.1). The check ring moves axially with the screw and rotates at a slower speed than the screw due to drag from a thin melt film between it and the barrel surface. 189

Mixing in Single Screw Extrusion

Figure 10.1 Injection moulding check ring. (Reproduced with permission from M. Gale in Proceedings of a Rapra Technology Ltd., Conference on Injection Moulding: Advanced Technology for Optimisation of Operational Performance, 1994, Shawbury, Paper No.4. ©1994, Rapra Technology)

Figure 10.2 Potential adaption for extrusion of the injection moulding check ring mixer by Elbe. 190

Floating Ring Mixing Devices

Figure 10.3 Vortex mixer with pins interacting with flights. (Reproduced with mermission from C. Rauwendaal, R. Maurer and M. Scheuber in Proceedings of the Annual SPE Conference – ANTEC, Nashville, TN, USA, 2003 p.183. ©2003, SPE)

Interacting elements can be incorporated providing the spacing accommodates the small axial displacement between the open and shut-off positions. Spacing of interacting pins in the ’turbine arrangements’ of Chapter 9 such that intermesh is avoided has been described by Elbe [3]. A schematic diagram by the author showing a possible adaption of Elbe’s pegged check ring mixer is shown in Figure 10.2. Slots or flutes resembling the outlet, placed at the mixer inlet might allow minimal relative axial ring movement and smaller clearances, but at extra cost. For good distributive mixing of colour masterbatches, wide spacings may be adequate. A check ring mixer using interacting pins in the stator ring has also been described by Rauwendaal for use in both extrusion [4] (see Figure 10.3) and injection moulding [5]. A mixer developed by Twente University [6-10], the Twente Mixing Ring (TMR) using this principle has been shown (like other interacting rotor/stator devices) to outperform more commonly used Maddock and pin mixers [11] (see Section 8.5.3). The rotor and stator arrangement is similar to the prototype cavity transfer mixer (CTM) in Figure 9.11. A complete injection moulding unit of this type is shown in Figure 10.4(a) and (b). A re-arrangement of the holes in the ring from a staggered pattern to separate rows can be used to strengthen the ring to withstand injection pressures. Lack of cavity alignment during injection is of no consequence as there is no mixing requirement at this point in the cycle, and there are no potentially intermeshing projections. An example of injection mouldings with a colour masterbatch, with and without such a device is shown in Figure 10.5(a) and (b). Note that the check ring 191

Mixing in Single Screw Extrusion

(a)

(b)

Figure 10.4 18 mm Injection moulding (TMR type) check ring cavity mixer. (a) Components, (b) Assembled.

(a)

Figure 10.5 View inside injection moulded caps with back lighting. (a) Without check ring mixer, (b) With check ring mixer. 192

(b)

Floating Ring Mixing Devices

Figure 10.6 Check ring mixer in forward and rear position with bore extended into nozzle.

mixer used was extended into the nozzle (Figure 10.6) making it longer than the original in Figure 10.1.

10.3 Adaption of the Check Ring Mixer to Extrusion By transferring the check ring mixer concept to extrusion, the mixer can be accommodated within the barrel as part of the screw. Being without the non-return valve function required for injection moulding it can be described as a ‘floating ring mixer’. As the stator is free to turn, it will rotate with the screw approximately half the screw speed or less, depending on the barrel clearance. Compared with a fixed stator configuration, for floating ring devices generally there is reduced mixing between the rotor and turning stator ring, but with the TMR, there is compensation for the reduced mixing from the rotor/sleeve interaction by the through cavities in the rotating sleeve interacting with the barrel surface (Figure 10.7). This will also reduce the risk of polymer stagnation in the sleeve/barrel gap, particularly if the cavity rows overlap (Figure 10.4). The advantages for extrusion is that there is no requirement to accommodate an extension between barrel and die, as will be the case for the interacting units described in Chapter 9. This is particularly useful for retrofitting to many blow moulding machines where splitting the clamp unit from the extruder is impracticable. There can be other situations where providing extra space between barrel face and die is difficult or expensive with existing extrusion installations. 193

Mixing in Single Screw Extrusion

Figure 10.7 Cross section of TMR type floating ring mixer showing flow patterns.

The disadvantage is that there will be less screw available for melting/pumping, and the length of the mixer is restricted. Although initial costs will be lower than for addon mixers, the need for occasional replacement of worn injection moulding check rings may also apply to extruders. A comparison of three floating ring mixers has been made by Myers and co-workers [12] using a 63.5 mm diameter, 21D (diameters) extruder with the screw length reduced to 19D to provide space for the mixers. The mixers compared were a TMR and two Barr mixers. The Barr sleeve mixer had elongated cavities in the rotor which joined pairs of circular cavities in the ring during rotation. The second cavity was connected to the next pair via a space between ring and barrel, the ring kept centrally by raised flanges between each diversion (Figure.10.8). The Barr ring mixer had five rotor rings equally spaced on a central shaft with 6 floating rings alternating with the rotor rings (Figure 10.9). With holes in both rotors and rings, polymer was subjected to a mechanism partly resembling a CTM, but at right angles rather than in-line, i.e., between discs rather than between cylinders. Experiments were performed using black and white acrylonitrile-butadiene-styrene pellets in a ratio of 220:1 white to black, with extruder speeds of 40 and 80 rpm. About 60% of the ABS entering the mixers was white with no black mixed in, which created big demands on the 2.4D mixers. Cross-sections of 1cm diameter strands showed the conventional screw had the most unmixed white sections, some of it unmelted. 194

Floating Ring Mixing Devices

Figure 10.8 Barr floating ring mixer. (Reproduced with permission from J.A. Myers, R.A. Barr, M.A. Spalding and K.R. Hughes in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 1, p.157. ©1999, SPE)

Figure 10.9 Barr multi-ring mixer. (Reproduced with permission from J.A. Myers, R.A. Barr, M.A. Spalding and K.R. Hughes in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 1, p.157. ©1999, SPE) 195

Mixing in Single Screw Extrusion The sleeve mixers were judged to perform well with minimal striations considering the amount of completely unmixed material entering the mixers. At a typical commercial 1:25 let-down ratio, mixing for the three mixers was adequate for acceptable extruded sheet. Trials with a Maddock mixer showed mixing to be considerably less than for floating ring mixers, confirming the conclusions of Esseghir and co-workers [11] presented in Section 8.5.3.

References 1.

M. Gale in Proceedings of a Rapra Technology Ltd., Conference on Injection Moulding: Advanced Technology for Optimisation of Operational Performance, 1994, Shawbury, Paper No.4.

2.

S.Y. Lin and M.J. Bevis, Plastics and Rubber Processing and Applications, 1987, 7, 1, 29.

3.

W.D. Elbe inventor; Vereinigung zur Foerderung des Instituts für Kunststoffverarbeitung in Industrie und Handwerk an der Rhein-westfalen Technical Hochschule Aachen, assignee; DE 2162709C3, 1974.

4. C. Rauwendaal, R. Maurer and M. Scheuber in Proceedings of the Annual SPE Conference – ANTEC, Nashville, TN, USA, 2003 p.183. 5. C. Rauwendaal, R. Maurer and M. Scheuber in Proceedings of the Annual SPE Conference – ANTEC, Nashville, TN, USA, 2003 p.526. 6.

G.J.M. Semmekrot, inventor; University Twente, assignee; EP. 0,340,873B1, 1992.

7.

University Twente, inventors; University Twente, assignee; NL 8801156, 1989.

8.

J.F. Ingen Housz, inventor; no assignee;US 4,218,146, 1980

9.

A.J. Ingen-Housz and S.A. Norden, International Polymer Processing, 1995, 10, 120.

10. British Plastics and Rubber, 1992, April, 26. 11. M. Esseghir, C.G. Gogos, Y. Dong-Woo, D.B. Todd and B. David, Advances in Polymer Technology, 1998, 17, 1, 1. 12. J.A. Myers, R.A. Barr, M.A. Spalding and K.R. Hughes in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 1, p.157. 196

11

Static (or Motionless) Mixers

As the term implies, unlike dynamic mixers described in previous chapters which turn with the screw, static mixers have no moving parts. These mixers are not specific to plastics extrusion but are versatile chemical engineering devices used for pipeline fluid mixing over a wide viscosity range. They can also be used for chemical reactions at very high flow rates. In plastics extrusion they have advantages of ease of installation with no requirement to be attached to, or incorporated with, the extruder screw. As they need to be positioned between the extruder output flange and the die, the requirement for extra space may place them with the same retrofit disadvantages as add-on dynamic mixers. However, with no screw connection required they can be installed within any available space between extruder outlet and die inlet, even at right angles to the axial line of the extruder barrel, e.g., vertically for a blown film extrusion line. The two main applications are improved cross-sectional distributive mixing of additives such as pigments and temperature homogenisation. This includes melt cooling, for example, in foam extrusion.

11.1 Mixing Mechanism Static mixers increase interfacial area by a combination of shearing and physical rearrangement. The polymer melt is passed through a tube containing a series of ducts. These are formed by groups of elements which combine shear flow with dividing and recombining the polymer passing through.

11.2 Static Mixers Used in plastics extrusion A number of different static mixers have been described together with the pressure losses which vary over a wide range between the highest and lowest [1]. Smith divided commercial static mixers into two categories of helical mixers and honeycomb packings [2]. 197

Mixing in Single Screw Extrusion Helical Mixers: Those that divide the flow between two or three channels and produce the re-orientation as a result of imposing a rotation on flow. Honeycomb Mixers: Those that provide flow re-arrangement with a packing allowing multiple open channels of flow.

11.2.1 Helical Mixers Four commercial mixers with quite different vane arrangements were included by Smith [2] with a summary of the Kenics mixer (Figure 11.1] regarded as the best known of this type for plastics extrusion. The example shown in Figure 11.2 is a plastic version used with a two component dispenser for mixing epoxy resin and hardener. 1) Flow is divided into two or more equal channels at the start of each element 2) A twist in the element of 90º or more turns the fluid along a helical path as it passes through the element 3) The following element reverses the turning direction The flow twisting arrangement of the Ross ISG mixer is completely different in that it uses two sets of crossing tubular ducts within each element to provide the redistribution, but this mechanism has a relatively high pressure drop compared with other mixers [1]. The Ross ISG and Kenics mixers and their mixing mechanisms have been described in some detail by Tadmor and Gogos [3].

Figure 11.1 Drawing of Kenics mixer (see Smith) [2]. (Reproduced with permission from J.M. Smith in Proceedings of a PRI conference – Polymer Extrusion 2, London, UK, 1982, Paper No.20. ©1982, Plastics and Rubber Institute. ©1982, PRI) 198

Static (or Motionless) Mixers

Figure 11.2 Photo of Kenics type static mixer.

11.2.2 Honeycomb Mixers Several honeycomb mixers have been developed by Sulzer/Koch which use packings typically consisting of corrugated sheets stacked in layers at right angles to each other and at 45º to the tube axis (Figure 11.3).These stacked layers form elements 1 to 1.5 D long and, as with the helical mixers, meet typically at 90º. Relative pressure drops range from the same as for Kenics mixers (which is otherwise the lowest in the comparison) to about average [1].

Figure 11.3 Drawing of Sulzer SMV honeycomb mixer. (Reproduced with permission from J.M. Smith in Proceedings of a PRI Conference - Polymer Extrusion 2, London, UK, 1982, Paper No.20, Figures 4a and 5. ©1982, Plastics and Rubber Institute) 199

Mixing in Single Screw Extrusion

11.3 Application in Heat Exchangers With no moving parts to generate heat, static mixers are adaptable for melt cooling heat exchangers, for example in foam extrusion. The division and re-orientation exposes polymer from the centre to the cooled outer surface. At the same time the design needs to be selected to minimise pressure drop as otherwise back pressure will cause more heat to be generated by the extruder screw.

11.4 Disadvantages There are a number of potential disadvantages in use, when compared to add-on dynamic mixers: 1) As this method of mixing inevitably involves pressure losses, this will impose a limit on the number of elements that can be used. For example, in a laboratory trial using a static mixer to mix carbon black masterbatch into low-density polyethylene the head pressures were so high that output rate dropped by 50%. Removal of half the elements restored output rate to a more acceptable level, but even with the extruder contributing extra mixing due to the increased back pressure, the extrusion failed to meet pipe standards (Figure 11.4). Fortunately, in many applications the required mixing is down to10-4 m, approximately the resolution of the human eye. 2) The flow resistance of the mixer can impose excessive stresses on the mixer itself. Following element folding, in an angled plate static mixer where the welded elements proved unable to withstand the loads applied in a melt cooling application, it was replaced by a single piece element machined from a solid steel blank. This proved suitable for the task (Figure 11.5). 3) The extra added volume requires more purging when changing materials, and the added residence time may contribute to thermal oxidation [1]. 4) Many of these mixers cannot be readily dismantled for cleaning, should this be necessary. 5) Most static mixers need treating with care to avoid damage. A lop-sided production line blown film bubble proved to be the result of heavy handed re-assembly of a static mixer following a routine total machine clean-down. 6) Static mixers used for plastics extrusion are relatively expensive [1]. 200

Static (or Motionless) Mixers

Figure 11.4 Photomicrograph of extrudate using Static Mixer (in MR or training course presentation/Powerpoint. (Reproduced with permission from G.M. Gale, Distibutive Mixing in Plastics Extrusion, Rapra Technology Members Report No. 46, Rapra Technology, Shawbury, Shrewsbury, UK, 1980, Figure 15. ©1980 Rapra Technology)

Figure 11.5 Drawing of Mixer used for cooling in foam extrusion.

201

Mixing in Single Screw Extrusion

References 1. C. Rauwendaal in Mixing in Polymer Processing, Ed., C. Rauwendaal, Marcel Dekker, New York, NY, USA, 1991, Chapter 4. 2. J.M. Smith in Proceedings of a PRI Conference - Polymer Extrusion 2, London, UK, 1982, Paper No.20. 3. Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley & Sons, New York, NY, USA, 1979.

202

12

Incorporation of Liquid Additives and Dispersions by Direct Addition

Many plastics additives are either liquids or solids which will melt at plastics processing temperatures. Some examples are shown in Table 12.1.

Table 12.1 Additives which can be incorporated as liquids into polymer melts Liquid additive

Applications

Liquid colour

Identification and decoration

Tackifiers/oils

Pallet and silage wrap, adhesives.

Lubricants (silicones)

Optical fibre ducting

Peroxides for crosslinking

Foams, cables, heating pipes

Silanes for crosslinking

Cables, heating pipes

Chemical foaming agents Physical foaming agents

Insulation, packaging, wood substitutes

Antioxidants UV absorbers

Heat resistance and outdoor exposure

Process aids

White oil in polystyrene (PS)

Crosslinking co-agents

Cables

In most cases the additives are available as masterbatches, (although liquid colours are essentially masterbatches of pigment concentrations in a liquid carrier). Direct addition can provide very significant economic and technical advantages, but usually accompanied by additional technical and safety responsibilities for the converter. There may be fire and/or toxicity hazards which need to be considered. Of particular concern in the above list are peroxides, silanes and triallyl isocyanurate crosslinking agent but others can be hazardous if heated at higher than manufacturers’ recommended temperatures or for too long, These include polybutenes, which are widely injected into extruders making tacky silage and pallet wrap film. 203

Mixing in Single Screw Extrusion

12.1 Viscosity Differences When mixing two viscous liquids their viscosities should ideally be the same [1]. When two liquids of widely different viscosities flow through a channel the situation may be as shown in Figure 12.1. Mixing a small quantity of a low viscosity additive into a high viscosity polymer melt will be less difficult than mixing a small quantity of high viscosity additive into a low viscosity polymer. There appears to be no rules for real situations [2], but it is not at all difficult to inject a liquid into the downstream regions of an extruder and produce a fine spray of the liquid around an extruded strand as it emerges from the die, e.g., white oil added to polystyrene. It is evident that the liquid can form a durable lubricating layer between the molten polymer and both barrel and die surfaces even when mixing elements are used. Fortunately there are ways of avoiding these problems.

Figure 12.1 Laminar shear flow with two liquid of different viscosities.

12.2 Incorporating Liquid Additives In principle a liquid additive can be added to polymers anywhere from the base of the hopper to an add-on mixer at the screw tip. In a few cases such as plasticisers with polyvinylchloride (PVC) polymer powder and hydrocarbon oil with styrene/butadiene thermoplastic elastomer pellets, batch premixing can be used, during which the polymer absorbs the liquid. The dry blend can then be processed by the extruder. For economic reasons, polybutene tackifier is frequently incorporated into linear low-density polyethylene (LLDPE) during extrusion of pallet wrap and silage wrap films. Taking this additive as an example, there are four points at which a liquid can be added (Figure 12.2): 204

Incorporation of Liquid Additives and Dispersions by Direct Addition a) Via the hopper into the feed throat. A low pressure pump can be used. A peristaltic pump is suitable for liquid colours. Screw conveying requirements limit addition rates to low levels b) Feed zone after the feed opening (point A). Pressures are quite low and gear pumps are used. Liquid addition levels are quite low before barrel slippage occurs, but higher than for feed opening addition. c) Down the centre of the screw to near the start of the metering zone (from B1 to B2). High pressure piston dosing pumps are used. Injection on to the screw surface causes less interference with screw pumping and higher polybutene levels can be incorporated. d) Highest incorporation levels are achievable by injecting into, or just before, a screw tip mixing device (point C), but this is the most complex and expensive process.

Figure 12.2 Polybutene injection points for silage and pallet wrap film extrusion.

By injecting between the screw tip and add-on mixer, or even into a first row cavity of a cavity transfer mixer, a poppet value injector can be used with a probe terminating well within the polymer melt. The poppet valve stops polymer backing up into the injection system (Figure 12.3). An injector with adjustable opening pressure is also possible. For high levels of liquid injection (see Figure 12.4) a number of actions can be taken to prevent surface films forming: 1) Ensure the injector tip protrudes well into the melt. 2) Have a number of axial troughs, webs and so on, around the section of screw/mixer opposite the injection probe. These can be formed by extending the entry channels to a cavity transfer mixer or injection can be into one of the first row stator cavities. 205

Mixing in Single Screw Extrusion 3) The liquid feed can be split between several injection ports spaced around the circumference. 4) A small extension probe can be fitted to the injector tip. The late addition of additives can have a number of benefits: 1) When changing colour, material wastage and changeover time will be reduced; a particularly important factor in wire insulation, due to the relatively high value of wasted copper wire. 2) Avoidance of hopper flow and feed zone conveying irregularities due to very sticky or very slippery masterbatch pellets. 3) Prevents premature heating effects such as peroxide crosslinking and gas escaping back through the hopper from a chemical foaming agent. 4) Technically innovative processes are possible.

Figure 12.3 Poppet valve liquid injector. (Drawing by M. Mackay, ©1988, Rapra Technology) 206

Incorporation of Liquid Additives and Dispersions by Direct Addition

Figure 12.4 Arrangement of poppet valve injectors to avoid slip layers. 207

Mixing in Single Screw Extrusion

12.3 Some Examples of Liquid Injection Processes 12.3.1 Polybutene in Pallet-wrap and Silage-wrap Film Polybutene is incorporated into polyolefines (frequently octene LLDPE) for these products. The film may have up to three coextruded layers with polybutene in two of them. This process has been well established for some years and data on such properties as the influence of polybutene on ultimate bond and rate of generation of tack properties have been produced by a polybutene supplier [3]. The level of addition, which ranges from 2% to 8.5%, depending on the application, has some influence on the injection point, which in turn influences the type and cost of pump required. Direct injection is carried out using special pumping equipment from either bulk storage tanks or individual drums, with preheating of the polybutene to 80 °C to reduce viscosity and hence reduce pressure losses through pipework and check valves etc. There are four points at which the polybutene can be added. This is illustrated in Figure 12.2. Injecting near the screw tip avoids any potential conveying problems at the feed zone, which may occur at high polybutene concentrations but requires expensive high pressure dosing piston pumps and flow rate monitoring as the polybutene is compressible.

12.3.2 Injection of Liquid Colours (General) Liquid colours are fundamentally the same as solid colour masterbatches. Both consist of a high concentration of pigment well dispersed in a compatible carrier. An advantage of liquid colours, which are pigments dispersed in a liquid carrier, is that they can be prepared in a similar way to inks and paints. The wetting of pigment by the liquid carrier minimises risks of agglomerate formation. This is particularly useful for wire insulation whether used directly or as a precursor for flexible PVC compound preparation. A single pigment agglomerate could cause electrical insulation failure. The liquid carrier in this instance is a normal PVC plasticiser. The main advantages of liquid colours are as follows: 1) The low viscosity liquid carrier facilitates rapid wetting of pigments to produce good dispersion. No melting is required and hence mixing is at ambient temperature. 2) Comparatively simple low powered (high speed) easily cleaned mixing equipment enables a fast customer response for colour matching. The fast response favours low inventories. 3) Liquid colours can be used with direct dosing into the polymer melt at, or near, the screw tip to reduce changeover times. It has been claimed that this arrangement 208

Incorporation of Liquid Additives and Dispersions by Direct Addition reduced purge times from 30 to 2 minutes, which at three times per day reduced purging waste by 90% [4]. 4) Proprietary handling, blending, and dosing equipment may well be available from the liquid colour supplier specifically for addition at the hopper (Figure 12.5).

Figure 12.5 Arrangement for liquid colour pre-blended with polymer granules for fibre extrusion. (Reproduced with permission from Colour Matrix. ©2008, Colour Matrix)

5) The system is overall cost effective, for example, for a translucent green pigmentation of a polyethylene terephthalate bottles, colour cost can be as little as 1.2% of overall material cost [5].

12.3.3 Wire Insulation Colouring An advantage of injecting liquid colour into a screw tip mixer for insulated wire extrusion is that time for colour change can be reduced typically at a 300 m/min production rate from 10-20 minutes to 3 minutes which results in a big financial 209

Mixing in Single Screw Extrusion saving [4] (mainly from reduced wastage of copper wire). Expensive pigment can be saved by confining the pigment to the surface of the insulation. For single wire extrusion, un-pigmented melt can be diverted into the cable crosshead to form the bulk of the insulation with the remainder being coated on to the surface following incorporation of pigment. This is achieved by passing the small proportion of melt (not bypassed) through a screw tip mixer where injected liquid colour is incorporated. This is essentially a co-extrusion using one extruder (Figure 12.6).

Figure 12.6 Cable co-extrusion using a single screw extruder with externally by-passed mixer. 210

Incorporation of Liquid Additives and Dispersions by Direct Addition Skin colouring, PVC wire insulation by mixing liquid colours into natural PVC just before the die has a number of advantages: 1) Liquid colour dispersions minimise the risk of pigment agglomerates being present. 2) Reduced costs from minimal quantity of expensive pigments being used. 3) Effect of pigment agglomerates on insulation is virtually eliminated by having no pigments present in the main body of the insulation. 4) Reduced costs from quicker changeover from one colour to another. The liquid colour injection equipment is shown in Figure 12.7 [6].

Figure 12.7 Colorant injection module. (Reproduced with Permission from D.W. South in the Rapra Technology Ltd, Symposium on Making More of the Cavity Transfer Mixer, 1986, Shawbury, Shrewsbury, UK, p.23. ©1986, Rapra Technology) 211

212

Wires

Radial (mm)

Outer diameter (mm)

7/1.04

7/1.35

6.0

10.0

1.0

0.8

0.7

-

-

6.0

10.0

1.2

1.1

0.9

3 core

3 core

5 core

5 core

6.0

35.0

1.5

6.0

1.5

1.4

1.7

1.5

1.4

16.10

10.60

27.0

13.9

9.6

8.7 × 17.25

103.20

60.69

202.67

87.65

54.09

106.00

82.61

37.71

24.45

15.07

6.92

Weight of insulation/ length of wire (kg/km)

1.0320

0.6069

2.0267

0.8765

0.5409

1.06

0.8261

0.377

0.254

0.1507

0.0692

Masterbatch – 1% (kg/km)

0.188

0.1088

0.316

0.1625

0.111

0.1653

0.1376

0.0865

0.0374

0.0279

0.0169

Liquid colour (kg/km)

5.4820:1

5.57:1

6.39:1

5.39:1

4.84:1

6.41:0

6.0:1

4.35:1

6.81:1

5.40:1

4.274:1

Ratio of masterbatch to liquid colour

42,200

24,905

85,535

35,700

21,495

44,731

34,425

14,500

10,830

6140

2615

Colour savings per 50,000 km (kg)

6242Y is a flat twin PVC cable to BS 6004 [8] NYM is a PVC sheathed general energy cable Reproduced with permission from K.P. Storton in the Proceedings of a Rapra Technology Ltd., Symposium - Making More of the Cavity Transfer Mixer, 1988, Shawbury, Shrewsbury, UK, p.9. ©1988, Rapra Technology

3 core

1.5

4.9 × 8.9

6.4

4.8

2.8

7.2 × 13.75

C Circular sheaths for NYM cables

-

1.5

B Flat twin and earth sheaths 6242Y

1/1.38

1.5

A Single core insulation PVC SG 1.5 HO7V

Size (wire cross section) (mm2)

Table 12.2 Comparison of pigment used for full colouring to that used by liquid skin colour

Mixing in Single Screw Extrusion

Incorporation of Liquid Additives and Dispersions by Direct Addition

Figure 12.8 Three die heads with three different colours using one extruder.

A system for producing mains electricity wiring with brown (live), blue (neutral), and yellow/green (earth), surface coloured insulation simultaneously is shown in Figure 12.8 [7]. The output from the main extruder is divided into separate streams for each wire, the flow rate for each individual stream being controlled by a gear pump. The main flow is then passed into the die head for un-pigmented insulation. A small proportion is bypassed via an individual mixer with liquid colour injection and mixing. The streams are then combined in the die heads by normal co-extrusion techniques to give unpigmented insulation with a co-extruded pigmented surface for three wires achieved with one extruder. An additional colouring unit could produce a coloured stripe, whilst a second extruder plus colouring unit supplied the sheathing. Full details of colour savings for different types and sizes of wires are shown in Table 12.2 [7].

12.3.4 Fibre Extrusion Similar considerations regarding both dispersion and quick colour changes in wire extrusion also apply to fibre extrusion. When combined with gear pumps, the dynamic 213

Mixing in Single Screw Extrusion mixers of the interacting cavity type will provide flexibility and efficiency required by demanding applications [9, 10] (See also Section 9.3.2).

12.3.5 Skin Colouring Pipes and Profiles The co-extrusion techniques described previously, for insulated wire can also be used for pipes and profiles, but the late incorporation of liquid additives also provides opportunities for novel techniques which are comparatively simple. For in-line extrusions, colour can be confined to the product surface by having a central bypass through the centre of a mixer and injecting the liquid colour after a blister which restricts flow into the mixing area (Figure 12.9). By re-uniting the streams at the mixer exit, a co-extrusion can be produced with additive confined to the skin. The result is that an in-line co-extrusion such as a pipe can be produced with only one extruder. Interchangeable rings can be used for the blister to give control of final thickness. Figure 12.10 shows a sample of 40 mm outside diameter (OD) skin coloured acrylonitrile-butadiene-styrene (ABS) pipe and the by-passed 38 mm rotor of the cavity transfer mixer (CTM) used for its extrusion. A typical bypassed mixer used for applying an expensive weathering protection additive in a pipe surface skin is shown in Figure 12.11(a) and Figure 12.11(b).

Figure 12.9 Co-extruded skin using mixer with central by-pass. 214

Incorporation of Liquid Additives and Dispersions by Direct Addition

Figure 12.10 CTM (38 mm) with central bypass and 40 mm ABS pipe with coloured skin. (Photograph bt G. Lawley. © Rapra Technology)

(a)

(b)

Figure 12.11 Production mixer with central bypass. (a) Shows entry slots before flow resisting blister, outlet pipe and liquid injector passing through stator flange; (b) View showing central bypass outlet. (Photographs bt G. Lawley. © Rapra Technology) 215

Mixing in Single Screw Extrusion

Figure 12.12 Three pumps spectrum system.

By having three pneumatically operated piston injection pumps, (ideally mounted directly to the three non-return valve injectors), there is an opportunity to apply an unlimited range of surface colours by adjusting the output settings of individual pumps (Figure 12.12). In a short demonstration trial it was shown possible to work through the spectrum from red to violet and also produce brown.

12.3.6 Crosslinking Polyethylenes Polyethylenes (PE) have been used for cables and pipes for cold water services for many years. They are comparatively easy to extrude with good thermal stability at about 150-180 °C for low-density polyethylene (LDPE) and LLDPE and about 175-200 °C for high-density polyethylene (HDPE). They are also competitively priced. However their applications are limited by a comparatively low upper service temperature limit. This applies to both continuous use under pressure at up to 100 °C for hot water pipes and occasional excursions to significantly higher temperatures supported by the conducting wire due to temporary electrical overload of a cable. 216

Incorporation of Liquid Additives and Dispersions by Direct Addition However, their safe working temperature can be raised to meet the demands of both products by crosslinking the polyethylene after the extrusion process has been completed. Although the same techniques are used for water pipes as for the longer established crosslinked cables, application of cable technology to pipes needs changes to satisfy the different service conditions. Other advantages of crosslinking polyethylenes include the following [11]: 1) Reduced deformation under load. 2) Improved chemical resistance. 3) Increased abrasion resistance. 4) Memory characteristics (for shrink sleeving). 5) Improved impact properties. One might add: 6) Enhancing foaming and foam properties.

12.3.6.1 Peroxide Crosslinking Peroxides dissociate into free radicals at a temperature specific to the selected peroxide. With the peroxide mixed into polyethylene a free radical reaction takes place which results in crosslinks between the molecular chains resulting in a network of molecules which increase the heat distortion temperature of the polymer. To achieve a crosslinked cable or pipe, the peroxide must be well mixed into the polyethylene in the extruder during normal extrusion, but the polymer must at all times be kept below the peroxide dissociation temperature. Failure to keep the polymer below this critical temperature will result in crosslinking within the extruder, producing at best a lumpy extrusion and in an extreme case blockage of the equipment. The application of screw cooling can help to limit melt temperatures as well as improve mixing [12-14]. The peroxide is normally pre-compounded into the polymer or added as masterbatch. Note: that excessive screw cooling can freeze polymer on to the screw, effectively reducing the screw channel working depth which increases melt temperature instead of cooling it. By injecting a liquid peroxide into the polymer melt between two Maddock elements at the screw tip or at the start of a suitable add-on mixer, higher temperatures can be safely used early in the extrusion process to accelerate polymer melting. The addon mixer described in [4] was preferred to a double Maddock arrangement which required eight injection ports spaced equally around the circumference. 217

Mixing in Single Screw Extrusion Following the extrusion process, insulated wire and cable can be heated to dissociate the peroxide into free radicals and crosslink the polyethylene. Pipe initially at its melting point and with no wire support will need special supporting facilities (which are normally proprietary) for post extrusion crosslinking. The injection system has been described as the same as that used for liquid colouring [4] but obviously there are safety issues associated with liquid peroxides which need to be considered before using this process.

12.3.6.2 Silane Grafting The basic chemistry of this process is that a silane such as tri-methoxy vinyl silane is grafted on to polyethylene and following extrusion the insulated wire or pipe is exposed to hot water or steam. The silane groups react with the water to form cross links with the elimination of methanol (Figure 12.13) [11]. A tin catalyst such as dibutyl tin dilaurate speeds up the crosslinking reaction. In the original Midland Silicones/Dow Corning process the grafting reaction is carried out in a compounding process and supplied packed in foil lined bags to prevent premature reaction with atmospheric moisture. The tin catalyst is supplied as a masterbatch. Crosslinking in hot water or steam enables pipes to be crosslinked below the polyethylene softening temperature, whilst there are no problems of crosslinking in the extruder, although the compound must be extruded immediately after bags are opened. A one step process was developed by BICC with Maillefer for power cables which eliminated the compounding stage (the Monosil process) [15], in which all the additives were mixed with the polymer and grafting performed in the cable extruder. Direct injection has been used to inject a liquid mixture of silane, peroxide and tin catalyst into an add-on mixer to make cable [4] and hot water pipe [16]. The water pipes were steam autoclaved for four hours at 110 °C, which was well below the softening point of the HDPE used and the resulting crosslinked pipes withstood 1000 hours in water at 95 °C with a wall stress of 4.4 N/mm2 [11]. As with peroxides there are health and safety issues. Tri-methoxy vinyl silane is both very flammable and toxic.

218

Incorporation of Liquid Additives and Dispersions by Direct Addition

Figure 12.13 Silane grafting and crosslinking reactions.

219

Mixing in Single Screw Extrusion

12.3.7 Silicone Lubricant Injection An example of liquid injection which gives both technical advantages and economic savings is a special application in which a silicone lubricant is incorporated at a level of 2% during the extrusion of small bore HDPE tubing. The purpose of the silicone is to provide a very low coefficient of friction. The tube is extruded to high tolerances with continuous laser monitoring. The addition of liquid silicone at the hopper would make it difficult to keep within tolerances, whilst silicone masterbatch would be expensive and may also give dimensional problems. By direct injection into a CTM added to the extruder, the required level of silicone is incorporated into the HDPE with no dimensional tolerance compromises and economic savings are made by purchasing the silicone in 25 kg drums instead of using a masterbatch.

12.3.8 Extrusion Foaming Extruded foams usually fall into one of the two categories of high density and low density. The former have densities typically of about half the original polymer density whilst the latter can be as low as 20-30 kg/m3. These two categories also subdivide the manufacturing processes, type of blowing agent and obviously the product markets. The high density foams, which are typically wood substitute products, generally use conventional extrusion lines and chemical blowing agents specially formulated to liberate the necessary gas. The most commonly used chemical blowing agents (CBA) liberate either carbon dioxide (endothermic type) or nitrogen plus miscellaneous gases, e.g., ammonia (exothermic type). A widely used example is azodicarbonamide. Low density foams, which are typically packaging and insulating products, are foamed on special extrusion equipment using direct injection of a physical blowing agent which does not undergo any decomposition reaction. Physical blowing agents are usually either hydrocarbons such as pentane or hydrofluorocarbons (HFC). Hydrocarbon blowing agents are of comparatively low cost, but flame-proof plants are required and liberation of hydrocarbons into the atmosphere is generally unacceptable. 220

Incorporation of Liquid Additives and Dispersions by Direct Addition 12.3.8.1 Extrusion Foaming Mechanisms The formation of a foamed thermoplastic has been described in articles by Throne [17] and by Han and co-workers [18]. The foam extrusion process can be summarised as: 1) Gas is introduced to the melt either via a CBA added to the feed pellets as powder or masterbatch or by injection of a physical foaming agent. The melt pressure developed by the extruder screw against the die restriction enables a solution of the gas in the molten polymer to be achieved. 2) As the die pressure reduces at or near the die exit, the solution becomes supersaturated and bubbles become nucleated at points of irregularity such as CBA solid residues, pigments, contaminants and deliberately added nucleating agents such as talc. 3) With further melt pressure reductions and transfer from die to atmosphere, the bubbles expand and the remaining gas comes out of solution to further expand the bubbles, forming a cellular extrusion. A number of factors will influence this mechanism: 1) As pressure is higher in small bubbles than larger ones, the larger bubbles will grow at the expense of the smaller ones. As a result it is desirable to nucleate as many bubbles as possible when using physical foaming agents. This can be achieved with foaming agents by adding an optimum concentration of sodium bicarbonate/citric acid mixture, talc and other fillers, silica and so on. CBA are self nucleating. 2) The degree of expansion depends on the amount of gas dissolved in the polymer. This will depend on the solubility of the gas and the melt pressure as described in more detail next. 3) The expansion of bubbles without foam collapse will depend on the melt strength. Thus amorphous polymers such as polystyrene have a comparatively wide processing range of satisfactory foaming temperatures, whilst semi-crystalline polymers such as polypropylene (PP) will have a very narrow range between solid and low viscosity fluid [19]. In the case of PP, special branched (viscous) grades are available. 4) Low density foams will be harder to achieve with gases which have a faster diffusion rate. 5) With dies designed to produce very rapid decompression at their outlets (e.g., by using wide entry angles and very short lands), transfer of dissolved gas from melt to expanding bubbles (as melt pressure rapidly drops) causes a corresponding 221

Mixing in Single Screw Extrusion rapid increase in polymer viscosity. This prevents bursting of the rapidly expanding bubbles thereby resulting in a lower foam density and also contributes to a more stable foaming process.

12.3.8.2 Conditions Required to Extrude Low Density Foams These can normally be achieved by taking advantage of plasticisation of the melt by the dissolved gas and adiabatic cooling by the gas expansion. Plasticisation enables the melt to be processed at a lower temperature, e.g., about 100 °C for polyethylene and polystyrene. As a result gas transfer from melt to bubbles during expansion causes the viscosity to rise rapidly. This is also assisted by adiabatic cooling of the melt by the expanding gas. Melt cooling to these extrusion temperatures can be achieved using either static heat exchangers or dynamically using a second (tandem) extruder with a deep screw running at a slow speed. Many CBA require high temperatures for decomposition and hence comparatively high melt temperatures are necessary at some stage.

Figure 12.14 Extruder arrangements for producing foam. 222

Incorporation of Liquid Additives and Dispersions by Direct Addition Figure 12.14 shows the individual extrusion stages: 1) The first stage is conveying and melting as for solid extrusions except that mixing must be adequate to achieve uniform distribution of the nucleating agent. 2) At the following mixing stage the foaming agent is injected under pressure into the molten polymer and well mixed into the polymer. The dissolved gas reduces the melt viscosity, i.e., acts as a plasticiser. 3) Cooling is applied at the third stage to restore the viscosity to its former or even a higher level to prevent cell wall rupture during subsequent bubble expansion as it leaves the die. Temperatures as low as 100 °C or even lower might be used for PS and LDPE. 4) The sudden pressure drop to ambient as the cooled melt leaves the die causes dissolution of the gas to form bubbles which very rapidly expand. It is evident that mixing is very important, not only to achieve a uniform distribution of nucleator and foaming agent, but also to achieve uniform melt temperature. With mixing producing heat, uniform melt cooling is not easily achieved.

12.3.8.3 Extrusion of Low Density Foams For low density packaging and insulation foam products, complete extrusion lines are available from several specialist plastics machinery manufacturers which may use one long single screw or twin screw extruder or two single screw extruders in tandem. (Figure 12.14). In the tandem process, the second extruder is designed specifically to combine extruder and heat exchanger functions. Due to the low thermal conductivity of polymers, the challenge is to expose all material to cooling surfaces without generating shear heat. As cooling increases viscosity, falling temperatures result in increasing generation of shear heating, which will stabilise temperatures at an equilibrium value.

12.3.8.4 Extrusion Foaming Using Chemical Blowing Agents For the high density end of the foam density spectrum, normal solid plastics extrusion machines can be used with chemical blowing agents either as a masterbatch or as a powder dusting on to the feed pellets. The output rate may be limited by the need to combine good melting and mixing whilst running the extruder at temperatures below which premature decomposition of the blowing agent and gas escape via the hopper 223

Mixing in Single Screw Extrusion occurs. One solution, which is not generally used, is to inject a liquid dispersion of blowing agent into a screw tip mixer, similar to the technique described for liquid colour in Section 12.3.2. The blowing agent and colour can be combined in the same carrier. This may minimise the variables in gas evolution associated with heating rate: a problem identified in [20]. The author found this technique worked well with a film blowing laboratory extruder and it also avoided gas escape via the hopper. Although carbon dioxide liberated from chemical blowing agents costs about 10 times that purchased in cylinders, the former offers advantages of simplicity of addition of powder or masterbatch pellets at the hopper. Overall, for ‘wood-like’ and similar extrusions, conventional single screw extruders can be used, as for solid extrusions.

12.3.8.5 Carbon Dioxide as a Foaming Agent There has been considerable research into the use of inert gases as replacement foaming agents for chlorofluorocarbons (CFC). This followed the range of legislation which followed the Montreal Protocol to eliminate the use of these chemicals which were believed to be damaging the ozone layer in the world’s upper atmosphere. hydrochlorofluorocarbons which were considered to be 90% less harmful than CFC were introduced as an interim measure, but these in turn were phased out in favour of HFC in 2001 [21]. The apparent overall advantages of nitrogen and carbon dioxide raises the question of why not use these gases directly as a replacement for both physical and chemical blowing agents? Nitrogen is by far the most environmentally acceptable contender as it is simply borrowed from the atmosphere. However, although carbon dioxide is a greenhouse gas, for foaming processes it is a by-product of ammonia manufacture, fermentation etc. Solubility in polymers is a critical property in achieving low densities. As this solubility is defined by Henry’s Law, i.e., the amount which will dissolve is proportional to the applied pressure, comparisons can be made using Henry’s Law constants. Comparisons shown in Table 12.3 show that three inert gas alternatives have significantly lower solubilities in PE, PP and PS than a CFC which can no longer be used. Carbon dioxide also has an advantage in that mixing into a polymer is aided by it being a critical fluid at normal extrusion temperatures and pressures such that it behaves like both a fluid and a gas, i.e., mixing is aided by diffusion. Table 12.4 summarises results of extrusion trials using injection of these inert gases into a cavity transfer mixer fitted to an extrusion line normally used for producing LDPE pipe insulation. Although a twin screw extruder was used the CTM rotor was fitted to only one of the screws, and followed by a static haet exchanger melt cooler and static mixer blender. A similar arrangement using a single screw extruder has been used on a laboratory scale with a 50 mm extruder [22]. 224

Incorporation of Liquid Additives and Dispersions by Direct Addition

Table 12.3 Comparisons of Henry’s Law constants (solubilities) for carbon dioxide, nitrogen and argon Henry’s Law Constant at 188 °C cm3/g atm (STP) Polyethylene

Polypropylene

Polystyrene

Nitrogen

0.111

0.133

0.049

Argon

0.133

0.176

0.093

Carbon dioxide

0.435

0.499

0.388

STP: Standard temperature and pressure

Table 12.4 Comparison of lowest achieved densities with theoretical densities Gas

Die Pressure (MPa)

Die Gas dissolved temperature (cm3/g at STP) (°C)

Theoretical density (kg/m3)

Actual density (kg/m3)

Nitrogen

7.2

118.71

3.628

109

200

Argon

7.0

120

6.5

96

200

Carbon dioxide

7.4

110

16.28

46.7

53

The extrusion processes used are fundamentally the same as those used for physical foaming agents. The densities achieved were quite low, but the lower density of the pipe insulation in production foamed with a hydrocarbon could not be achieved. A further disadvantage was that unsightly surface wrinkling appeared during the first hour after extrusion, although it had only marginal effect on foam density. This was a consequence of the carbon dioxide diffusing out of the foam faster than air/nitrogen could diffuse in at this comparatively low density. Advantages of using carbon dioxide are: 1) Environmentally acceptable. 2) Reasonable solubility. 3) Can be used for a wide range of processing temperatures. 4) Plasticising effect reduces melt viscosity. 225

Mixing in Single Screw Extrusion 5) Lower processing temperature increases solubility. 6) Behaves as a critical fluid under typical extrusion conditions (diffusion makes mixing easier). 7) Can be pumped as a liquid. 8) Conventional extruders easily adapted. 9) Low cost (but pumping equipment needed and expertise to operate it). 10) With no limitation of specific decomposition temperature, (as with readily available chemical blowing agents), carbon dioxide can be used for foaming at a range of temperatures covering many polymers. Table 12.5 shows results for several engineering plastics ranging from polypropylene to polyether-ethersulfone [23]. 11) No gas escape via the hopper.

Table 12.5 Foaming conditions, foam density and polymer density Polymer density (kg/m3)

Foam density (kg/m3)

Extruder barrel temperature (°C)

Heat exchanger/die temperature (°C)

Die type

Die Pressure (MPa)

PEEK

1300

350

390

280/320

Strip

5.52

PPS

1300

690

340

295/275

Strip

31.0

SMA

1140

123

190

140/130

Rod

6.34

HTPC

1200

374

340

250/250

Strip

24.1

Modified PPO

1100

208

280

200/200

Rod

6.76

Branched PP

900

150

190

105/140

Strip

4.55

Linear PP

900

360

190

120 (CTM)

Strip

1.65

Material

HTPC: High temperature polycarbonate PEEK: Polyether ether ketone PPO: Polyphenylene oxide PPS: Polyphenylene sulfide SMA: Styrene-maleic anhydride Reproduced with permission from G.M. Gale in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000, p.1945. ©2000, SPE

226

Incorporation of Liquid Additives and Dispersions by Direct Addition Disadvantages of using carbon dioxide are: 1) It must be injected under high pressure. 2) Liquid carbon dioxide is compressible which can make accurate dosing variable such that stable operating conditions may be difficult to achieve if dosing equipment does not allow for this. 3) Its limited solubility (compared with hydro- and fluorocarbons) restricts its use to high, medium and medium-low density foams. 4) It is generally unsuitable for foamed packaging trays, thermal insulation etc. which usually have very low densities. 5) A small chiller with trace cooling is necessary to keep carbon dioxide as a liquid in the pump and immediate pipework.

References 1.

J.M. McKelvey, Polymer Processing, John Wiley & Sons, London, UK, 1962.

2.

Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, New York, NY, USA, 1979.

3.

J.H. Gardner in Proceedings of a Rapra Technology Ltd., Conference Addcon ‘99, Prague, Czechoslovakia, 1999, Paper 8.

4.

A. Mattila in the Proceedinsg of a Rapra Technology Ltd., Symposium Making the most of the Cavity Transfer Mixer, Shawbury, Shrewsbury, Shropshire, 1986, Paper No.8.

5.

A. Overend, ColorMatrix (private communication)

6.

D.W. South in the Rapra Technology Ltd, Symposium on Making More of the Cavity Transfer Mixer, 1986, Shawbury, Shrewsbury, UK, p.23.

7.

K.P. Storton in the Proceedings of a Rapra Technology Ltd., Symposium Making More of the Cavity Transfer Mixer, 1988, Shawbury, Shrewsbury, UK, p.9.

8.

BS 6004, Electric Cables - PVC Insulated, Non-Armoured Cables for Voltages up to and Including 450/750 V, for Electric Power, Lighting and Internal Wiring, 2006. 227

Mixing in Single Screw Extrusion 9.

F. Hensen, Advances in Polymer Technology, 1984, 4, 3, 339.

10. F. Dickmeiss, Extrusion, 2007, 2, 20. 11. G.M. Gale, Applied Organometallic Chemistry, 1988, 2, 1, 17. 12. E. Steward in Proceedings of the Annual SPE Conference – ANTEC 2001, Dallas, TX, USA, 2001, Paper No.62. 13. B.H. Maddock, SPE Journal, 1967, July, 23. 14. D. Grant, and W. Walker, British Plastics, 1951, 24, 8, 308. 15. P. Swarbrick, W.J. Green and C. Maillefer, inventors; BICC Ltd and Establissements Maillefer SA, assignees, US 4,117,195. 16. A. Sorio and G.M. Gale in Proceedings of the PRI International Conference Polymer Extrusion 3, London, UK, 1985, Paper No.27. 17. J.L. Throne in Proceedings of the MIT International Conference on Polymer Processing, Cambridge, MA, USA, 1977, p.77. 18. C.D. Han, Y.W. Kim and K.D. Malhotra, Journal of Applied Polymer Science, 1976, 20, 1583. 19. J.G. Burt, Journal of Cellular Plastics, 1978, 14, 6, 341. 20. G.L.A. Sims and C. O’Connor in Proceedings of a Rapra Technology. Ltd., Seminar - Blowing Agent Systems: Formulation and Processing, Shawbury, Shrewsbury, UK, 1998, Paper No.2. 21. M. Jeffs, Urethanes Technology, 2004, 21, 6, 32. 22. H. Benkreira, M. Gale, R. Patel, M. Cox and J. Paragreen, International Polymer Processing, 2004, 19, 2, 111. 23. G.M. Gale in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000, p.1945.

228

13

Dispersive Mixing of Fillers and Pigments

13.1 Formation of Agglomerates Extruded products frequently contain additives in the form of powders ranging in concentration from less than 0.5% pigment or antiblock to 40 or 50 wt% filler or flame retardant. With moves to reduce use of halogens, water-liberating flame retardants such as alumina trihydrate whose water of crystallisation is liberated at about 170 °C, may be required at up to 50 wt% to meet specific fire retarding performances. There are attractions, either economic or technical (or both), in compounding particulate blends of polymers with these additives during product extrusion. Certain organic pigments will form exceptionally large agglomerates (termed comprimates by Wiese [1]) when they are compressed, as will happen during single screw processing. Practical trials showed that these comprimates needed a high level of energy for dispersion not available during normal processing conditions. For pigments, antiblocking additives, and other additives required at low concentrations, the masterbatch route is normally technically the best and most reliable. It is also the best for minimising cleaning and facilitating automatic and accurate dosing. For high levels of filler, the economics of the direct route look attractive, particularly for extrusions requiring high stiffness. Unfortunately, as pointed out by Smith [25], the compounding of solid additives into plastics involves situations that are the opposite of those ideally required for good dispersion (see Chapter 2). From the photograph in Chapter 1, it appears that some of Smith’s statements (and Wiese’s comprimates) apply in this particular case (although these agglomerates were filler, not pigments): 1) (There is) ‘... abundant evidence to show that agglomerates are normally produced under circumstances largely overlooked and unsuspected’. 2) (It is) ‘... largely true that once formed, an agglomerate stands a reasonable chance of never meeting conditions which break it down, the production of good quality dispersions is made difficult or impossible’. 229

Mixing in Single Screw Extrusion 3) ‘Problems of agglomeration can be overcome simply by avoiding their formation’. As discussed in Chapter 2, the overall problem in dispersive mixing, whatever machinery is used, is that compounding of solid additives into plastics involves ‘situations that are the opposite of those ideally required’.

13.2 Formation of filler agglomerates in a single screw extruder

The problems of extruding polymer pellet/filler powder blends in single screw extruders were investigated by examining samples of materials removed from the screw channel of a 38 mm diameter, 24:1 single screw extruder following rapid cooling and screw jacking [6]. The first trial used a conventional 3:1 depth ratio screw with equal length, feed, compression, and metering zones. The second trial used a screw with a Maillefer type barrier section replacing the conventional compression zone. Blends of 40 wt% stearate coated calcium carbonate filler with polypropylene (PP) pellets were prepared in a 24 litre high speed mixer run at half speed of 1200 rpm under vacuum to prevent fluidisation and also remove any moisture present. Butyl stearate (0.1 wt%) was included to minimise filler separation during extruder feeding. Mixer discharge temperature was 110 °C. Figure 13.1 shows PP granules coated with filler.

Figure 13.1 Polypropylene pellets coated in filler. 230

Dispersive Mixing of Fillers and Pigments

Figure 13.2 Sample removed from extruder screw channel.

Figure 13.2 shows a sample removed from the conventional screw channel at an early compression zone stage. The lower surface, which had been in contact with the screw, had patches of filler on it. During removal, the lightly bonded granules forming a stratum across the middle have separated showing little adhesion between pellets. The surfaces of the exposed pellets of the cross-section are coated with filler, and agglomerates have formed between pellets. From examination of all the screw channel samples, the following observations were made (Figure 13.3): 1) In the early stages of compaction, where melting should have started, a solid’s mass had formed, firstly against the barrel surface, and then spreading inwards to the full channel depth within several screw turns. The solid material appeared to be a sintered mass of polymer and filler with a polished surface skin which, in some places, had large patches of segregated filler as a surface coating. 2) It appeared that during passage through the compression zone the usual melting behaviour of a rolling melt pool, forming at the rear of the screw channel and steadily growing in width until the channel was full, did not occur.

Instead, the separation of polymer granules from one another by a coating of filler, resulted in polymer granules deforming under the effects of heat and pressure without coalescing. This appeared to transfer high compression forces to the powder filling the pellet interstices so that strong filler agglomerates were formed. Some of them were very large, i.e., Wiese’s comprimates. 231

Mixing in Single Screw Extrusion 3) Coatings of filler between polymer and metal surfaces prevented the melting polymer from wetting the barrel and screw surfaces. 4) The sintered mass continued through most of the compression zone in this compacted form, so that the pockets of filler must have been subjected to high compression. 5) At some point, which could not be clearly identified, the polymer in the sintered mass received sufficient conductive heat to become fully molten and normal crosssection circulating flow occurred. 6) The agglomerates were entrained within the melt through the die into the extrudate.

Figure 13.3 Formation of agglomerates. 232

Dispersive Mixing of Fillers and Pigments Figure 13.4 is a view of the channel cross-sections and screw contacting surface for the Maillefer type barrier screw. Note that the central strip formerly occupying the gap between the barrier flight and barrel surface has small patches of compressed filler and agglomerates passing over the barrier flight. The overall effect was that thermally softened filler coated PP pellets had deformed into platelets during passage over the barrier flight with coalescence still being prevented by the filler coating (Figure 13.5). Agglomerates already formed were also passing through the gap into the melt channel. The fitting of a Maddock mixer, planetary gear mixer, or cavity transfer mixer would not disperse the agglomerates.

Figure 13.4 Sample removed from barrier screw (Maillefer type).

Figure 13.5 Agglomerates passing over screw barrier flight. 233

Mixing in Single Screw Extrusion

13.3 Starved Feeding to Avoid Agglomerate Formation The investigation discussed in Section 13.1 confirmed Smith’s advice that the best way to avoid the presence of agglomerates is to avoid their formation in the first place. The requirement is therefore to forward polymer granules and additive powder with minimal compaction forces. This has led to several investigations into the use of starved feeding, i.e., controlling the rate of addition in the extruder feed zone such that the screw channel is less than full. The reasoning for this approach is that forwarding will rely on gravity conveying instead of particulate compaction, which is necessary to achieve sufficient drag forces from barrel surface friction for plug flow conveying. Transport of heavy materials such as sand and gravel is very easily achieved using this method (Figure 13.6). For polymer/additive blends, this method of conveying has the potential for moving pellets not only without compaction, but including a tumbling movement within each screw compartment [7]. This should uniformly raise the polymer/additive temperature as it is forwarded throughout the hot barrel towards the die (Figure 13.7).

Figure 13.6 Solids conveying mechanism. 234

Dispersive Mixing of Fillers and Pigments

Figure 13.7 Tumbling of powder/pellets in partly filled screw.

Finally, under the influence of die back pressure, polymer will become fully melted during passage though the last few turns of the screw. In fact this idea was not completely original, but a recognition that this starved feeding behaviour illustrated by Maddock [8] in screw jacking experiments could be suitable for the avoidance of agglomerates generated by normal full channel extrusion. This influence of starved feeding on melting and mixing was just one of many situations included in trials with colour pellet mixtures by Maddock in 1959 to investigate overall melting and mixing mechanisms [8]. When the feed rate was reduced to 50% of normal delivery rate, the length of filled channel was reduced to 4.5 turns from the discharge end, with improved mixing resulting from increased residence time in the metering zone. With the reduced feed rate unchanged, the screw speed was then doubled. This reduced the length of filled channel still further, but improved mixing more than before. The mixing improvement was attributed to the increase in total shear, i.e., mixing is directly related to shear rate multiplied by residence time. Further trials demonstrated that restricting output rate with a valve or die could be used to control onset of melting. 235

Mixing in Single Screw Extrusion In 1978, McKelvey [9] speculated that the extra degree of freedom provided by decoupling extrusion rate from screw speed might lead to unrecognised advantages. His trials with a 203 mm extruder confirmed that with starved feeding the screw channel was part filled for most of its length, with full width being reached after completion of melting. However, the main interest was potential improvements in process economics. It was shown that higher overall output rates and energy savings should be possible (particularly for large extruders with very viscous styrene-acrylonitrile). Of particular relevance to more recent work using starved feeding for avoiding agglomerate formation, was McKelvey’s observations on surging occurring as a consequence of starved feeding. Varying output rates even with an accurate dosing system is a potential problem with starved feeding of single screw extruders. One might expect starved feeding would eliminate risks of variable output rate (or ‘surging’). It should replace potentially variable friction dependent solids feeding and full screw conveying with a controlled feeding device and ‘gravity conveying’ in the screw. Unfortunately, as shown by McKelvey (and experienced by the author), starved feeding can result in surging. A potential cause is down channel progress of polymer being slowed (or stopped) by failure of polymer melting to keep pace with reducing channel depth in a conventional screw. Once the delaying material has become fully melted, down channel flow will resume until more un-melted material arrives to repeat the cycle. If starved feeing is producing this effect, then controlled melting using a barrier screw might eliminate this problem. Both Thompson and co-workers [10] and Elemans and van Wunnik [11] carried out starved feeding trials using barrier screw extruders with the objective of achieving good dispersion of additive powders. The materials used and the additive levels to be dispersed were quite different. Between them, they covered the typical commercial extremes of a high level of filler in a commodity polyolefine (Thompson and co-workers), and a low level of pigment in an engineering polymer (Elemans and van Wunnik). Thompson and co-workers used 20 wt% calcium carbonate filler with high-denisty polyethylene (HDPE) pellets and black marker pellets. Elemans and van Wunnick used a dry coloured blend of polybutylterephthalate (PBT) with 2% ultramarine blue pigment premixed by tumble blending. The extruders were of a similar size, but the one used for the HDPE/filler trials was longer to accommodate a vent zone and had an alternative conventional screw. The PBT/pigment trials used a screw with a mixing element following the barrier section. As the two investigations used screws of similar diameters and together covered a range of screw features, the essential details are compared in Table 13.1. 236

Dispersive Mixing of Fillers and Pigments

Table 13.1 Summary of screw details used in starved feed zone extrusion trials by Thompson, Donoian and Christiano [10] and by Elemans and Van Wunnik [11] Mixer

Feed zone depth (mm)

After barrier depth (mm)

Yes

No

12.7

3.3

No

Yes

No

12.7

–

Yes

No

Yes

11.2

3.4

L/D

Diameter (mm)

Barrier

Vent

HDPE/filler barrier screw

31D

63.5

Yes

HDPE/filler conventional screw

31D

63.5

PBT/pigment barrier screw

28D

60

With the HDPE/filler feed, a particular feature for the starved feed condition was the significant reduction in melting zone pressure which coexisted with reduced agglomeration. This could have been the result of pellet preheating causing softening prior to compaction which as a result, reduced compaction forces during melting. Overall, Christiano and co-workers reported mixing was better under starved conditions with both barrier and conventional screws; the agglomerate size reducing with increase in starvation, with little difference between general purpose and barrier screw. The most significant improvements were achieved using up to 10% starvation. With the conventional screw, a consequence of 15% starvation at a higher screw speed was surging, accompanied by a re-appearance of agglomerates. In Elemans and Van Wunnick’s trials with PBT/2% ultramarine blue, extrudates and injection moulded plaques showed significantly fewer agglomerates when screw speed was increased from a flood fed 40 rpm condition to a starved 60 rpm at the same feed rate of 40 kg/h (Figure 13.8). Further screw speed increases gave no further improvement. Starved feeding by reducing the feed rate (as for the HDPE/filler) showed the same improvements, although compared with the high filler content situation, increasing under-feeding from 70% to 45% further improved dispersion. These results (Figure 13.9) can be correlated with pressure measurements from points between 12D and the die entry (Figure 13.10). 237

Mixing in Single Screw Extrusion

Figure 13.8 Extrusions comparing flood-feed with starved feed by increasing screw speed at constant throughput rate. (Reproduced with permission from P.H.M. Elemans and J.M. Van Wunnik, Polymer Engineering and Science, 2001, 41, 7, 1099. ©2001, Wiley)

Figure 13.9 Photomicrograph showing reduced agglomerates as feed rate reduced at constant screw speed. (Reproduced with permission from P.H.M. Elemans and J.M. Van Wunnik, Polymer Engineering and Science, 2001, 41, 7, 1099. ©2001, Wiley)

Figure 13.10 Graph showing lower pressure resulting from increasing starvation at constant speed. (Reproduced with permission from P.H.M. Elemans and J.M. Van Wunnik, Polymer Engineering and Science, 2001, 41, 7, 1099. ©2001 Wiley) 238

Dispersive Mixing of Fillers and Pigments

13.4 Dispersive Mixing Using Polymer Powders The best dispersions of ultramarine blue in PBT achieved by Elemans and van Wunnick, were achieved by using PBT in powder form. The polymer powder gave a 100 fold increase in surface area compared with pellets over which the pigment can be distributed. This gave a considerable improvement over the use of granules and with starved feeding, no agglomerates were readily discernable [11] (Figure 13.11).

Figure 13.11 Photomicrographs showing pigment dispersion improved by using starved feeding with polymer powder. (Reproduced with permission from P.H.M. Elemans and J.M. Van Wunnik, Polymer Engineering and Science, 2001, 41, 7, 1099. ©2001, Wiley)

If the polymer is not available as reactor powder and needs pellet grinding, the costs will probably be uneconomic. If reactor powder is used, there may be a risk of its availability being discontinued as happened with PP reactor powder used in an experimental pin barrelled extruder. In this case, a flood fed 20 mm extruder with four pins of 3 mm diameter mounted radially at 90° to each other and at 2 D intervals, passed through slots in the screw flights [6]. Good dispersion was achieved with PP powder with 20 wt% chalk filler blends, whereas dispersions were poor when PP granules were used. Soon afterwards, the reactor producing the PP powder was closed down, and at the time there was no alternative supplier. However, in the production of colour masterbatches, linear low-density polyethylene reactor powder is sometimes used.

13.5 Dispersive Mixing Using Polymeric Waxes A technique enabling single screw extruders to achieve good dispersion of pigments in colour masterbatches has been describe in the past by polymer wax manufacturers. 239

Mixing in Single Screw Extrusion An advantage of this process is that a number of small single screw extruders which are of relatively low cost, can economically produce small orders for specialised colours. They can be easily cleaned between batches or maybe dedicated to a single colour. The attraction of this approach has probably declined over more recent years following the introduction of very small twin screw compounding extruders. As explained in Chapter 2, the medium into which pigments are to be dispersed should ideally have a low viscosity such that it can easily wet powders and penetrate between particles, preventing agglomeration should small particles be pressed together during mixing. The reality with plastics extrusion, is that the opposite situation is normally the case. Polymeric binders can be selected to be polar or non-polar to match the pigment polarity, be compatible with the main polymer constituting the masterbatch carrier, and melt to a low viscosity liquid which provides the pigment wetting. A range of such waxes can be found in polymer wax suppliers’ literature [12]. As the level of pigment in a masterbatch can be 25-50% or higher, the wax level, which might adversely affect the final moulded product’s qualities, will, at possibly 15% of the masterbatch, be at a very low overall concentration. Some viscosity reducing matrices can also be favoured to improve distributive mixing during screw processing [13]. An example taken from a Luwax brochure [12] is as follows: Formulation: 25% Ultramarine blue

15% Luwax A or Luwax AL3



60% PE or PP granules

Mixing is carried out in a high speed/cooler combination mixer, similar to the type used for polyvinylchloride (PVC) dry blend powder compound preparations (Figure 13.12). During mixing, the temperature rises with time as shown in Figure 13.13: 1) The polymer wax melts to become a low viscosity liquid which wets and coats the pigment particles 2) About 5 °C-10 °C above the wax melting point, the wax/pigment mixture forms a coating on the plastics granules. 3) The blend is discharged into a slow speed mixer or into stirred separate easily cleanable or dedicated containers. 4) The cooled blend is compounded into pellets using a single screw extruder. 240

Dispersive Mixing of Fillers and Pigments

Figure 13.12 Single screw masterbatch compounding using high speed mixing/ wax techniques.

Figure 13.13 Behaviour of wax/pigment blend during high speed mixing.

For ‘easy to disperse’ pigments, the level of wax can be reduced. For example:

50% titanium dioxide 10% chalk 5% Luwax (HDPE wax)* 34% low-density polyethylene granules 1% antioxidant

*These are polymeric waxes suitable for styrenics, PVC and polyamide. 241

Mixing in Single Screw Extrusion

References 1.

V.K. Weise in Proceedings of the Annual SPE Conference – ANTEC, Atlanta, GA, USA, 1975, p.143.

2.

M.J. Smith, Journal of the Oil and Colour Chemists’ Association, 1973, 56, 3, 126.

3.

M.J. Smith, Journal of the Oil and Colour Chemists’ Association, 1973, 56, 4, 155.

4.

M.J. Smith, Journal of the Oil and Colour Chemists’ Association, 1973, 56, 5, 165.

5.

M.J. Smith, Journal of the Oil and Colour Chemists’ Association, 1974, 57, 1, 36.

6.

G.M. Gale in Proceedings of the Annual SPE Conference – ANTEC, Montreal, Canada, 1991, p.95.

7.

G.A. Campbell, G. Nagarajan, J.D. Small and J. Staples in Proceedings of the Annual SPE Conference – ANTEC, Boston, MA, USA, 1995, p.268.

8.

B.H. Maddock, SPE Journal, 1959, 15, 5, 383.

9.

J McKelvey in Proceedings of the Annual SPE Conference – ANTEC, Washington, DC, USA, 1978, p.507.

10. M.R. Thompson, G. Donoian and J.P. Christiano, Polymer Engineering and Science, 2000, 40, 9, 2014. 11. P.H.M. Elemans and J.M. Van Wunnik, Polymer Engineering and Science, 2001, 41, 7, 1099. 12. Luwax Brochure, BASF Performance Chemicals, Ludwigshafen, Germany. 13. H. Benkreira and R.N. Britton, International Polymer Processing, 1994, 9, 3, 205.

242

14

Dispersive Mixing Applied to Polymer Blending

Single screw extruders generally give poor dispersive mixing and the blending of polymers is theoretically a dispersive mixing process requiring elongational shear as for the dispersion of pigments and agglomerates [1-3]. There are numerous published articles examining many polymer combinations whilst two books on the subject consist of several volumes [4, 5]. In most cases internal mixers or twin screw extruders are used for compounding polymer blends. However, as much of the fundamental work, which dates back to the 1930s uses simple laminar shear flow models and Couette flow, it is apparent that there should be a niche for single screw extruders. This raises the question of: how large is this niche?

14.1 Polymer Blends In practical terms, polymer blends can be put into three categories: 1) Combinations which give properties that are better than might be expected from their individual properties. 2) Combinations which offer a predictable but useful balance of properties at an economic cost. 3) Useful materials which can result from homogenising difficult to separate mixed polymer scrap and waste. For this situation the terminology is as follows: Scrap defines process scrap produced during product extrusion plus any additional converting operations such as edge trim and skeletal scrap from thermoforming. Co-extrusion scrap will become a polymer blend. Other materials will be re-used as appropriate as single polymers. Waste defines post consumer materials such as used packaging and ‘end-of-life’ products such as computer housings. Complete (100%) separation may not be economic. 243

Mixing in Single Screw Extrusion In terms of morphology we can divide blends into two categories: 1) Miscible Blends: Those which are readily compatible and can be easily combined by distributive mixing. They will usually have properties which are predictable from their component properties and blend ratio. These blends can generally be identified by retaining transparency of the original polymers and by having a single glass transition point. There are a number of exceptions, and it is also possible to have a reversible change from miscible to immiscible with the same blend as the temperature is raised [6]. 2) Immiscible Blends: These polymer blends have separate phases, much like an oil in water emulsion with the minor phase suspended as droplets in the continuous major phase (or matrix). The majority of polymer blends are in this category. With the addition of materials to provide compatibilisation, polymer blends can provide properties which are better than might be predicted from those of the individual components. Without compatibilisation, mechanical properties may be poor. An early example of enhanced properties is high impact polystyrene (HIPS) in which an elastomeric polymer exists as tiny droplets within a comparatively brittle continuous phase. This increases impact strength by restricting crack propagation. New materials have been produced from existing polymers by using combinations where the advantages of the one compensates for the disadvantages of the other. An example is a blend of polyamide-6 (PA6) or polyamide-66 (PA66) with polypropylene (PP) [7]. This blend has a higher service temperature than PP and a lower water absorption than polyamide (PA). Such blends are usually proprietary compounds produced by polymer manufacturers to specific technical specifications as for a single polymer. The polymerisation process may also play a part. Polycarbonate (PC)/acrylonitrile-butadiene-styrene (ABS) represents a typical blend that from its commercial introduction in the 1960s has become a major material in the automotive industry, as well as applications in business machine and electronic sectors. The blends provide the balance of properties required for these applications at an economic cost. These include low temperature ductility/impact and processability including injection moulding, extrusion and thermoforming. The commercial growth has been maintained by developments in heat stability, thermal ageing in hot wet environments and halogen free flame retardency [8]. Blends of polyethylene (PE) or PP with polystyrene (PS) have lower mould shrinkage and increased hardness compared with PE and improved stress cracking resistance compared with PS [9]. PS has a higher viscosity than PE and consequently dispersing the minor phase will be easier when it is PE and harder when it is PS, but this will also be influenced by a number of variables as explained below. 244

Dispersive Mixing Applied to Polymer Blending Conventional mixing of PE with PS for a broad compositional range will give poor mechanical properties if the separate phases are not held together with a compatibiliser. Higher tensile strength and low elongation at break will result from the PS and yield stress will be lower than either alone as a result of poor interfacial bonding. The phases will easily debond except at very low temperatures [5]. By using suitable diblock copolymer emulsifiers (compatibilisers), Fayt and co-workers [10] produced laboratory samples of some interesting polymer combinations such as: 1) PS/PE (superior weathering to HIPS) 2) Polymethylmethacrylate/styrene acrylonitrile (better impact than ABS) 3) Polyvinylidene fluoride/Noryl (improved physico-mechanical properties) 4) PS/PA6 (range of very different properties) Although most polymers are (surprisingly) incompatible as shown in Table 14.1 [11], they are blended during film extrusion to meet specific properties such as stiffness, film thickness etc., within cost restraints, although co-extrusion can sometimes achieve superior properties at lower cost [12].

Table 14.1 Example of a miscibility guide Polymer

PS

PA

PC

PVC

PP

LDPE

HDPE

PS

Y

PA

N

Y

PC

N

N

Y

PVC

N

N

N

Y

PP

N

N

N

N

Y

LDPE

N

N

N

N

N

Y

HDPE

N

N

N

N

N

N

Y

PET

N

N

N

N

N

N

N

PET

Y

Y = miscible; N = immiscible HDPE: High-density polyethylene LDPE: Low-density polyethylene PET: Polyethylene terephthalate PVC: Polyvinylchloride Reproduced with permission from V. Goodchild, Introduction to Plastics Recycling, 2nd Edition, Smithers Rapra Technology, Shawbury, Shrewsbury, UK, 2007, Table 4.1.

245

Mixing in Single Screw Extrusion

14.2 Polymer Scrap With the need for improved barrier properties to meet food shelf life requirements, extruded films and sheet may have five layers or more using three or more polymers. With the generation of up to 40% skeletal scrap from thermoforming plus extruded sheet edge trim, re-use will be necessary for the process to remain economic. This may require homogenisation by a ‘buried layer’ extruder. This machine may need a specific mixing screw to homogenise the scrap sufficiently to avoid ripples and other surface defects. Scrap may also be extrusions with dimensional or appearance defects which can be reused blended with virgin polymer at an agreed ratio. Providing various housekeeping and quality issues are met, there should be no mixing difficulties, but there are instances where a separate extrusion line processing 100% scrap with suitable screw mixing facilities is necessary.

14.3 Polymer Waste In recent years, much of the research into polymer blending has focussed directly or indirectly on recycling post consumer waste. Much of this is packaging consisting mainly of polyolefines but also significant amounts of PS, and PET. An audit [13] is summarised in section 14.9. From this data we can be confident that the recycling of polymer waste will include mixing of immiscible materials. For example: 1) PS is immiscible with polyolefines, 2) PP is immiscible with PE, 3) Linear low-density polyethylene is immiscible with LDPE, but miscible with HDPE ([14] see also Table 14.1). The perceived need for dispersive mixing for polymer blending raises the question as to how single screw extruders, which are poor dispersive mixers, can produce useful products where the material to be recycled consists of polymer mixtures. However, as happens so often in extrusion, nothing can be taken for granted.

14.4 Blending Immiscible Viscous Fluids Mixing two incompatible polymers together is regarded as a dispersive mixing process [1-3]. The dispersed phase of the blend exists as tiny droplets of a few microns 246

Dispersive Mixing Applied to Polymer Blending diameter within the continuous phase, and it is assumed from fundamental theory that they originate from large droplets which need to be broken down into smaller ones. There is therefore an analogy with pigment and filler dispersion in which dispersive forces break agglomerates down into their ultimate particle size. There are also many differences. The situation of a droplet within a continuous phase has been defined as follows [15]: • ‘When one liquid is at rest in another immiscible liquid of the same density, it assumes the form of a spherical drop. Any movement of the outer fluid will distort the drop because of the dynamic and viscous forces which act on its surface, Interfacial tension however will tend to keep the drop spherical’. Other factors include: • The higher the viscosity of the continuous phase, the greater the ease of break-up of the liquid drop. • The smaller the drop, the higher the shear rate needed to break it. • The more a drop is deformed, the more unstable it becomes and hence it tends to break up. Experiments carried out to develop the droplet dispersion theory have mainly used apparatus applying simple laminar shear flow, i.e., the mechanism existing in single screw extrusion. Most papers refer to work in the 1930s by Taylor [16] who carried out model experiments observing droplets in laminar shear fields produced by two parallel belts moving in opposite directions, as subsequently used by Theodorou for agglomerates described in Chapter 2. Karam and Bellinger [15] used Couette flow (also described in Chapter 2) with an annulus formed by counter-rotating concentric glass cylinders. Taylor’s work with Newtonian liquids showed that elongation of a droplet is favoured by low interfacial tension, larger particle diameter, matrix viscosity and high shear rates. Flumerfelt [17], using non-Newtonian fluids, showed that in a simple shear field, a spherical drop becomes ellipsoidal with the major axis and inclined at about 45º from perpendicular to the shear field. Depending on relative viscosities, a critical shear rate was reached in which the droplet broke up into smaller droplets. A minimum size was eventually reached below which break-up could not be achieved regardless of shear rate. 247

Mixing in Single Screw Extrusion

Figure 14.1 Droplet in laminar shear field.

Droplet break-up was less likely to exist where it had either a relatively high or relatively low viscosity. In the latter case the inclined droplets had pointed ends where very small drops broke away in a stream. These were unlike the relatively high viscosity droplets which showed an ellipsoid shape nearly aligned with the flow direction (Figure 14.1). As mentioned in Chapter 12, when mixing two polymers together by laminar shear, their viscosities should ideally be similar. Furthermore, mixing small quantities of a high viscosity material into a low viscosity polymer is more difficult than mixing a small quantity of a low viscosity polymer into a high viscosity polymer. This also applies to the break-up of droplets. According to Taylor, the breaking up of a single drop requires the viscous forces acting on the droplet to exceed the interfacial forces for a sufficient amount of time. This also indicates that a limit may exist in single screw extruders whereby under certain conditions, laminar shear flow fails to achieve a critical level of stress needed to break up the droplet. 248

Dispersive Mixing Applied to Polymer Blending The requirement that for droplet deformation the tension of the deforming matrix (continuous phase) must overcome the interfacial tension, can be expressed as a ratio termed the Weber number (We) where:

We = Where

Tension of the matrix µmγ = Interfacial tension s /R

γ = rate of shear or elongation of the matrix µm = the viscosity of the matrix s = the coefficient of the interfacial tension R = radius of the undeformed droplets

Critical values for Weber numbers at which droplet break-up occurs have been determined using the moving belt, roller and Couette flow techniques (Chapter 2). By plotting critical Weber numbers against viscosity ratio (ρ), curves of the shape shown in Figure 14.2 are produced:

ρ=

µd µm

where ρ = the viscosity ratio µd = the dispersed phase viscosity µm = the matrix viscosity

Figure 14.2 Critical Weber number versus viscosity ratio. (Reproduced with permission from E. de Jong in a Rapra Seminar – Making More of the Cavity Transfer Mixer, Rapra Technology, Shawbury, Shrewsbury, UK, Paper No.15. ©1988, Rapra Technology) 249

Mixing in Single Screw Extrusion

14.5 Polymer Blending Mechanisms in a Single Screw Extruder The theory considers the minor phase of a blend starting as a larger droplet suspended in the matrix or continuous phase, rather like a pigment or filler agglomerate. In a similar manner to dispersing a pigment into smaller particles, the larger droplets are broken down into smaller droplets, but by using either laminar shear or extensional flow. Referring to Figure 14.2, the extensional/hyperbolic flow behaviour covering a wide viscosity ratio indicates the suitability of internal mixers and twin screw compounding extruders for reducing the droplet size of the dispersed phase. The graph for laminar shear flow, which will apply to single screw extruders shows a cut-off point where the ratio of disperse phase viscosity to continuous phase viscosity reaches a value of about 4. Following the increasing viscosity (from left to right) particle break-up will occur following the mechanism of C and then B (in Figure 14.1) with condition A occurring at a ratio approaching 4. Overall this indicates that a polymer blend of small dispersed droplets in a continuous phase cannot be produced using a single screw extruder if the viscosity ration of the disperse phase to the continuous phase is greater than 4. Practical experience indicates that this cannot be true. There is also the question as to how droplets are formed in the first place? Scott and Macosko [18] noted conclusions by others that the most significant changes in polymer blend morphology occurred during the first few minutes of mixing when polymer softening and melting occurred [19]. For their investigation into this behaviour, they carried out polymer blend mixing studies using a Haake Rheocord torque rheometer (a laboratory batch mixer which continuously records drive torque and melt temperature during mixing). Blends were prepared using 20wt% amorphous PA as the disperse phase in PS. Viscosity ratio of PA:PS was about 14; well above the limit of about 4 for laminar shear mixing. Samples were mixed for times varying from 1 to 15 minutes and examined by scanning electron microscopy (SEM) following solvent extraction of the PS. There appeared to be some similarity with the mixing on melting feature in single screw extruders described in Chapter 7, noting also the very high torque recorded during the first minute, falling quickly as the material temperature rapidly increased. After only one minute (which included about 27 seconds loading time) the materials ranged through several mixing stages: 1) Unmelted pellets. 2) Areas of sheets and ribbons - some with holes. 250

Dispersive Mixing Applied to Polymer Blending 3) A lace structure. 4) Spherical particles 0.5-3 µm diameter. Within another half minute there were many particles resulting from disintegration of the lace structure and by seven minutes nearly all disperse phase material was spherical particles. The results overall indicated the following mechanisms: • Dragging of pellets by the rotors across the mixer wall formed sheets or ribbons of the dispersed phase within the matrix. • Due to interfacial tension, instability caused holes to form. • When holes reached sufficient numbers and size, a fragile lace structure was formed which fell apart. • The lace fragments so formed were of a similar diameter to the particles generated at a later mixing time. • Breakdown continued until all fragments became spherical. When considering that the start of the droplet formation route in Scott and Makosco’s investigation was the conversion of pellets into ribbons by the torque rheometer it is no surprise that a single screw extruder can produce polymer blends using the same series of mechanisms. This was demonstrated in experiments by Willemse and co-workers [20] using blends of 5 and 17.5 wt% PS in PE extruded in a conventional 20 mm diameter extruder. This was fitted with a Ross ISG static mixer with one element to provide a stretching, folding and cutting mechanism (‘bakers transformation’ [21]), followed by up to 10 neutral tubular elements which extended residence time at constant shear. Morphology of the extrudate structures was examined and measured using SEM following solvent extraction of the PS. Pellets (3 mm) were reduced to a dispersed phase sheet thickness of 2 µm by the extruder as predicted by calculation. A further reduction to 0.2 to 0.8 µm occurred after the single static mixer element; with very little further change over the remaining nine neutral elements. These could be fibrils or droplets. Huong and Li [22] demonstrated the transformation from striation to droplet using a blend of PP, PA6 and compatibiliser in the ratio 85:12:3 wt%. 251

Mixing in Single Screw Extrusion A 30:1 L/D single screw extruder was used to compare three different screw designs: 1) Conventional 2) Conventional with addition of a fluted element about 2D long, positioned 18D from the feed end. 3) Conventional screw with addition of a pineapple/pinned element fitted in the same position. Materials were sampled via bleed ports before and after the mixer and at the screw tip. Overall the fluted element was the most effective and the plain screw the least in transforming striations into droplets. The final droplet sizes were approximately of the same order at 2.8, 2.1 and 2.6 µm.

14.6 Break-up of Fibrils into Droplets This is not necessarily a requirement as the threads may be an effective reinforcement. This behaviour has been comprehensively reviewed [23], whilst the mechanism for commercial polymer combinations has been clearly illustrated in experiments by Elemans and co-workers [24] from observations of droplet formation from filaments used to measure interfacial tension. The procedure was to sandwich extruded threads between two films placed between glass slides and observe under a microscope during heating to an appropriate temperature. (Figure 14.3). LDPE, HDPE and PA6 threads in PS film were examined. Depending on viscosities, filament thickness, and interfacial tension, with the same thread diameter of 20 µm, the time required ranged from 50 seconds for PA6/PS to several hours for PS/LDPE. Adding a diblock copolymer (compatibiliser) to the thread phase decreased interfacial tension for HDPE thread in PS at 200 °C from nearly 5 mN/m to a constant value of one over a range of 1-5 wt% (Figure 14.4). This increased the stability of the molten thread, as also observed by Yu and co-workers [25]. In addition to experiments with an extruder having a single static mixer element and neutral elements described previously, Willemse and co-workers [20] used the Ross ISG static mixer with its full compliment of 11 elements fed with two separate melt streams of PE and PS. Results showed that 2000 µm sheets reduced to 1 µm after 6 elements. Thereafter they remained as 0.3-0.4 µm particles. Results were very similar for both 5 wt% and 17.5 wt% PS. Overall, they concluded that the critical thickness for sheet break-up was not dependent on the viscosity of the matrix. Whether the threads broke up or not depended on the 252

Dispersive Mixing Applied to Polymer Blending

Figure 14.3 Stages in break-up of a PA6 thread embedded in polystyrene. (Reproduced with permission from P.H.M. Elemans, J.M.H. Janssen and H.E.H. Meijer, Journal of Rheology, 1990, 34, 8, 1311.)

Figure 14.4 Influence of compatibiliser on surface tension. (Reproduced with permission from P.H.M. Elemans, J.M.H. Janssen and H.E.H. Meijer, Journal of Rheology, 1990, 34, 8, 1311.)

capillary number (which is the same as the Weber number in Section 14.4). A fibre matrix was formed if the capillary number was >1, and a droplet if it was 1 such as PS, PVC, polyethyleneterephthalate The separation of PP from LDPE is not a viable process. 4) The presence of PP (commonly 10%) will make the LDPE stiffer but reduces both elongation at break and impact strength. 260

Dispersive Mixing Applied to Polymer Blending This investigation showed compatibilisation will significantly improve elongation at break and Charpy impact strength of samples compression moulded from extrudates prepared with a twin screw extruder. Results for uncompatibilised blends prepared in a 35 L/D single screw extruder with a high compression screw and mixing device running at 35 rpm were compared with those for a twin screw extruder 32 L/D run at 35-50 rpm. They concluded that (with LDPE/PP, 90/10) the twin screw extruder produced better tensile properties and more homogeneity. From the results, the elongation at break for virgin materials and for post consumer materials were 28% and 17.2%, respectively, and were greater for the twin screw extruder than for the single screw extruder. Other mechanical properties were reasonably similar. In judging polymer blending performance in single screw extruders, the important role played by screw speed should also be considered. Joshi and co-workers [41] found an optimum screw speed was dictated by shear rates and residence time. Although no firm conclusions concerning the suitability of single screw extruders for processing mixed polyolefin waste can be drawn from these two small scale extrusion investigations, they indicate that single screw extruders need to be designed specifically for recycling. Specialised screws and mixing devices may be necessary to compensate for the lack of flexibility of screw configuration and feed rate independence available on twin screw extruders. The direct extrusion of post consumer mixed polyolefin waste into film using standard film blowing plant would provide significant economic benefits.

14.10 Elongational Flow Mixing There will probably always be an interest in achieving elongational flow mixing in single screw extruders to provide good dispersion without incurring the capital costs of twin screw extruders. It is recognised that polymer melts will experience shearing during passage over the barrier in the Maddock and other fluted elements, but that a single or minimal number of passes is very restrictive. Although the addition of screw shearing elements and add-on devices are unlikely to satisfactorily disperse pigments and fillers agglomerates, there is a generally continuing interest in fitting elongational mixing devices to provide dispersive mixing for polymer blending and elimination of gels which can often appear during film extrusion. Two quite different types of static mixer have been devised. Gramann and co-workers [42] used parallel bars arranged in successive rows at right angles, and Song [43] used round plates forming concentric annular ridges and troughs (Figure 14.7). These provided 261

Mixing in Single Screw Extrusion

Figure 14.7 Radial static dispersive mixer. (Reproduced with permission from W. Song, in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000 p.270. ©2000, SPE)

alternating restrictive gaps and channels. Polymer entered around the periphery and left via a central hole. A cross section diagram [40] shows that flow from the extruder is through a central feed to the entrance around the periphery via a tube die, spider type arrangement. Trials with HDPE/PS (10:90 wt%) showed fibrils became droplets during passage through the mixer and that impact strengths for ethylene-propylene rubber/PP blends were superior to blends produced by twin screw extrusion. This type of mixer had also enabled a single screw extruder achieve better dispersion than a twin screw extruder (the latter with or without the mixer) when producing PA6/clay nanocomposites [44, 45]. A mixing element by Rauwendaal [46] using a combination of multi variable clearance flights and flights with slots was aimed at providing the required flow fields for dispersive mixing. Distributive mixing was achieved by re-arrangement between repeated dispersion stages. Microscopic examination of an extruded strand cross section produced from a HDPE/PS (60:40 wt%) pellet blend showed a domain size gradient from 2 µm in the central region to 20 µm in the outer third region. The coarser particles were attributed to droplet coalescence [47]. In mixer comparisons this Chris Rauwendaal Dispersive mixer (CRD) generally outperformed the others, but in all cases variations occurred with screw speed with no clear trends [48]. In a wide range of single screw mixers reviewed by Schut [49], the designs and applications were very diverse. The potential for a reduction in gels in extruded films was of specific interest. The results reported for the CRD mixer were varied, but this may have been due to the type of gels present (see Section 14.11).

14.11 Elimination of Gels Gels usually appear as tiny unsightly spheres about pin head size and are often found in packaging films. Their appearance can be magnified by their disturbance 262

Dispersive Mixing Applied to Polymer Blending to polymer flow in the die which may also cause lines which in severe cases reduce mechanical properties such as tear strength. They can also be mistaken for carbon black or pigment agglomerates. They may be scattered, appear as ‘gel showers’ at irregular intervals, or may accumulate on the die face (die drool) [50-52]. A short but concise article by Waller [53] reviewed the many types of gels, causes and remedies. From an extrusion mixing aspect, gels can be divided into two main types: 1) Those that might be dispersed by mixing. These (sometimes termed ‘unmelts’) may be higher molecular weight particles formed as a result of the polymerisation process. 2) Those that are unlikely to be dispersed by mixing. These are most likely crosslinked as a result of thermal oxidation. Before installing new screws or mixing devices to cure a gel problem, it is worth establishing which type of gels are present. As with agglomerates, the best approach if they are crosslinked is to avoid their formation in the first place. Oxidative crosslinking increases with exposure to high temperatures and long residence times, e.g., as a result of stagnation (‘hang-up’) in corners of adaptors, dies, etc. The following are some factors which have little direct relevance to mixing but should be considered in film extrusion: 1) Are melt temperatures exceptionally high? 2) Are there ‘dog legs’ in film co-extrusion feed pipes? 3) Does the extruder and die experience extended periods at a high temperature between production runs? Adding antioxidant before shut down reduces risks of generating gels of oxidised polymer during extended cooling and reheating. Screen packs will sieve out some gels, but having crosslinks enables gels regain their original spherical shape after being squashed through the mesh. Confirmation of crosslinks is possible using infra-red analysis. Following confirmation, the addition of antioxidant masterbatch should solve the problem (at increased material cost) providing good distributive mixing occurs.

References 1.

Z. Tadmor and C.G. Gogos, Principles of Polymer Processing, John Wiley, NewYork, NY, USA, 1979. 263

Mixing in Single Screw Extrusion 2.

H.P. Grace, Chemical Engineering Communications, 1982, 14, 3-6, 225.

3.

H. Potente and J. Flecke, in Proceedings of the Annual SPE Conference – ANTEC, Indianapolis, IN, USA, 1996, Volume 1, p.178.

4.

Handbook of Polymer Blends and Composites, Eds. A.K. Kulshreshtha and C. Vasile, Rapra Technology Ltd., Shrewsbury, UK, 2002-2003, [4 volumes].

5.

Polymer Blends Handbook, Ed., L.A. Utraki, Kluwer Academic Publications, Dordrecht, The Netherlands, 2002. [2 Volumes]

6.

O. Olabisi, L.M. Robeson and M.T. Shaw, Polymer-Polymer Miscibility, Academic Press, New York, NY, USA, 1979.

7.

J. Girard in Proceedings of a Rapra Technology Seminar – Engineering with Blends and Alloys, Shrewsbury, UK, 1989, Paper No.4.

8.

B. Hager, D. Wittmann and E. Weng in Proceedings of the Annual SPE Conference-ANTEC Milwaukee WI USA 2008, p.1301

9.

B.S. Munteanu, M. Brebu and C. Vasile in Handbook of Polymer Blends and Composites, Volume 4A, Eds., A.K. Kulshreshtha and C. Vasile, Rapra Technology, Shawbury, Shrewsbury, UK, 2003, Chapter 4.

10. R. Fayt, R. Jerome and P. Teyssie , Polymer Engineering and Science, 1987, 27, 5, 328. 11. V. Goodship, Introduction to Plastics Recycling, 2nd Edition, Smithers Rapra, Shawbury, Shrewsbury, UK, 2007. 12. J. Taylor and J.J. Baik in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000, p.209. 13. Waste on Line, www.wasteonline.gov.uk. Their original source was MEL, Aston Science Park, Birmingham, UK and RECOUP, Woodston, Peterborough, UK. 14. V. Musil, B. Pregrad and B. Zerjal, International Polymer Processing, 1988, 2, 3-4, 182. 15. H.J. Karam and J.C. Bellinger, Industrial and Engineering Chemistry Fundamentals, 1968, 7, 4, 2020. 16. G.I. Taylor, Proceedings of the Royal Society, Series A, 1934, 146, 501. 264

Dispersive Mixing Applied to Polymer Blending 17. R.W. Flumerfelt, Industrial and Engineering Chemistry Fundamentals, 1972, 11, 3, 312. 18. C.E. Scott and C.W. Macosco, Polymer Bulletin, 1999, 26, 341. 19. M.A. Haneault, M.F. Champagne, L.E. Daigneault and M.H. Dumoulin in Proceedings of the Annual SPE Conference – ANTEC, Boston, MA, 1995, Volume 2, 2020. 20. R.C. Willemse, E.J.J. Ramaker, J. van Dam and A. Posthuma de Boer, Polymer, 1999, 40, 24, 6645. 21. Y. Germain, Materiaux et Techniques, 1991, 79, 11-12, 3. 22. H-X. Huang and X-J. Li in Proceedings of the Annual SPE Conference ANTEC Cincinnati, OH, USA, 2007, p.337. 23. J.J. Elmendorp in Mixing in Polymer Processing, Ed., C. Rauwendaal, Marcel Dekker, New York, NY, USA, 1991, p.17. 24. P.H.M. Elemans, J.M.H. Janssen and H.E.H. Meijer, Journal of Rheology, 1990, 34, 8, 1311. 25. D-W. Yu, M. Esseghir and C.G. Gogos in Proceedings of the Annual SPE Conference – ANTEC, Boston, MA, USA, 1995, p.136. 26. C.G. Gogos, M. Esseghir, B. David, D.B. Todd, D.R. Sebastian and R. Garritano in Proceedings of the Annual SPE Conference – ANTEC, New Orleans, MA, USA, 1993, p.1542. 27. D. Herridge and D. Krueger, in Proceedings of the Annual SPE Conference – ANTEC, Montreal, Canada, 1991, p.633. 28. H. Aref. Journal of Fluid Mechanics, 1984, 143, 1. 29. L. Erwin, Polymer Engineering and Science, 1978, 18, 258. 30. G.A. Campbell, S. Bomma, S. St. John and S. Chempath in Proceedings of the Annual SPE Conference - ANTEC, San Francisco, CA, USA, 2002, Paper No.152. 31. S.C. Jana in Proceedings of the SPE Annual Conference - ANTEC, San Francisco, CA, 2002, Paper No.151. 265

Mixing in Single Screw Extrusion 32. O. Kwon and D.A. Zumbrunnen, Journal of Applied Polymer Science, 2001, 82, 1569. 33. M. Sau and S.C. Jana in Proceedings of the Annual SPE Conference ANTEC, San Francisco, CA, USA, 2002, Paper No.150. 34. D.A. Zumbrunnen and B. Kulshreshtha and A. Dhoble, in Proceedings of the Annual SPE Conference – ANTEC, Boston, MA, USA, 2005, p.238. 35. V.A. Chougule and D.A. Zumbrunnen, in Proceedings of the Annual SPE Conference – ANTEC, 2003, Nashville, TN, USA, p.1299. 36. V.A. Chougule, R.M. Kimmel and D.A. Zumbrunnen, in Proceedings of the 63rd Annual SPE Conference – ANTEC, Boston, MA, USA, 2005, p.2976. 37. B. Kulshreshtha, A. Dhoble and D.A. Zumbrunnen in Proceedings of the 63rd Annual SPE Conference – ANTEC, Boston, MA, USA, 2005, p.2418. 38. C. Mahesha and D.A. Zumbrunnen in Proceedings of the 64th Annual SPE Conference – ANTEC, Charlotte, NC, USA, 2006, p.491. 39. T. Kallel, V. Massardier-Nageotti, M. Jaziri, J-F. Gerard and B. Elleuch, Journal of Applied Polymer Science, 2003, 90, 28, 2475. 40. S. Bertin and J.J. Robin, European Polymer Journal, 2002, 38, 11, 2255 41. J. Joshi, R.L. Lehman and T.J. Nosker in Proceedings of the 63rd Annual SPE Conference – ANTEC, Boston, MA, USA, 2005, p.2097. 42. P. Gramann, B. Davis, T. Osswald and C. Rauwendaal, in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 1, p.162. 43. W. Song, in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000 p.270. 44. L.A. Utracki and G.Z-H. Shi in Polymer Blends Handbook, Ed., L.A. Utracki, Kluwer Academic Publishers, Dordrecht, Netherlands, 2003, Chapter 9. 45. L.A. Utracki, M. Li and J. Sepehr, International Polymer Processing, 2006, 21, 1, 3. 46. C. Rauwendaal, T. Osswald, P. Gramann and B. Davis, in Proceedings of the Annual SPE Conference – ANTEC, Atlanta, GA, USA, 1998, Volume 1, p.277. 266

Dispersive Mixing Applied to Polymer Blending 47. C. Rauwendaal, A. Rios, T. Osswald, P. Gramann, B. Davis, M. del P Noriega and O.A. Estrasa, in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 1, p.167. 48. J.H. Shut, Plastics Technology, 2005, 51, 7, 45. 49. J.H. Shut, Plastics Technology, 1999, 45, 7, 36. 50. N. Petiniot, European Plastics News, 1997, 24, 2, 33. 51. J.D. Gander and A.J. Giacomin, Polymer Engineering and Science, 1997, 37, 7, 1113. 52. C-M. Chan, International Polymer Processing, 1995, 10, 3, 200. 53. P. Waller, Plastics Technology, 2003, 49, 12, 36.

267

Mixing in Single Screw Extrusion

268

15

Compounding with Single Screw Extruders

From the numerous technical publications covering investigations into nano-materials, grafting reactions, recycling and so on, it appears that the compounding machines normally used are co-rotating twin screw extruders. For production compounding, the situation is different. A bar chart by Black [1] shows that more single screw extruders than twin screw are being used for production compounding, Single screw extruders have advantages of lower cost, simplicity, ruggedness, ease of maintenance, and the ability to combine compounding with product extrusion. They lack the dispersive mixing performance and the considerable flexibility of screw configuration, feed rate/screw speed decoupling, and multiple ports for venting and solids addition. Even so, there is interest in single screw innovations to improve their compounding attributes [2] whilst screw L/D ratios have been increased up to 50:1 operating at 800 rpm [3]. Six examples are described next which indicate the diversity of single screw compounding applications which have been evaluated, including several comparisons with twin screw extruders. In some cases, there is an inference that the single screw extruder machine’s attribute of being easier and cheaper to maintain, is a consideration. This applies to compounding materials which are not ‘extruder friendly; e.g., abrasives. 1) A 60 mm single screw extrusion line has been described by Sigl and Fritz [4], which produced long glass fibre reinforced polypropylene (PP) sheet (glass mat reinforced thermoplastic), by a single pass compounding/sheet extrusion process. This ‘single pass’ process produced thermoformable sheet with the high impact properties required by the automotive industry. To minimise both fibre damage and machine wear, the fibres were added to molten polymer via a downstream port into a decompression zone. A similar technique for incorporating filler was illustrated by Todd [5]. 2) Comparisons were made by Utracki and Sepehr [6] between single and twin screw extruders for compounding nano-clays into polyamide-6 and PP, in which the elongational flow mixer (EFM) described in Chapter 15, was used. The combination of single screw extruder plus EFM produced better dispersions than were achieved with a twin screw extruder with or without an EFM. 269

Mixing in Single Screw Extrusion 3) Stasiek [7] found that a co-rotating twin screw extruder was more effective than a single screw machine for compounding a range of mineral fillers into polypropylene, in part because of the former’s control of the mixing process, i.e., it avoided the compacting/agglomerating effect described in section 13.1

Hoekstra and co-workers [8] compounded crushed glass into high-denisty polyethylene (HDPE) to evaluate whether it could be used as an alternative to glass fibres and glass spheres to increase tensile and flexural modulus of recycled HDPE used as ‘plastics lumber’. Comparisons were made with wood flour, glass fibre and calcium carbonate. Compounds were prepared with a 25 mm single screw extruder for injection moulding into test pieces. The crushed glass did not perform as well as other reinforcements commonly used. Results showed minimal influence on tensile strength and notched impact for all four ingredients. For flexural modulus, the coarser crushed glass had virtually no effect, whilst the finer glass increased stiffness by almost 5% (similar to calcium carbonate at 5.3%). The wood flour was significantly better at 10.4% increase, whilst glass fibres provided the highest increase at almost 55%.

5) Kuan and co-workers [9] made comparisons using two roll milling, twin screw compounding and single screw extruder compounding of wood flour into polyvinylchloride. The best mechanical properties were obtained with the twin screw extruder. Similarly, with 30% wood flour/HDPE composites, tensile strengths were 18% greater for a twin screw extruder than with a single screw machine. 6) A single screw extruder and a twin screw extruder were compared by Kowalska [10], for the production of PP and LDPE compounds containing commingled comminuted rubber crumb from scrap tyres. A coupling agent was also incorporated. It was found that the maximum amount of rubber scrap that could be incorporated was 75% by weight for both polymers in the single screw extruder. The maximum for the twin screw extruder was 30 wt%, and it was considered inadvisable to use this machine as partial decomposition of the rubber occurred. A situation that often exists, which may be particularly relevant to waste recycling, is that deficiencies may appear with larger continuously running production extruders which were not apparent in the laboratory.

References 1.

T. Black in Plastics Compounding: Equipment and Processing, Ed., D.B. Todd, Hanser Publishers, Munich, Germany, 1998, Chapter 2.

2.

J.H. Schut, Plastics Technology, 1999, 45, 3, 46.

270

Compounding with Single Screw Extruders 3.

F.R. Pranckh, Plastics Compounding and Extrusion, 1998, 21, 6.

4.

K-P. Sigl and H-G. Fritz in Proceedings of the Annual SPE Conference – ANTEC, New York, NY, USA, 1999, Volume 2, p.2709.

5.

D.B. Todd, Advances in Polymer Technology, 2000, 19, 1, 34.

6.

L.A. Utracki, M. Sepehr and J. Li, International Polymer Processing, 2006, 2, 1, 3.

7.

J. Stasiek, Polimery, 2005, 50, 11-12, 881. [in Polish]

8.

N.L. Hoekstra, D.P. Duffy and S.H. Dillman in Proceedings of the Annual SPE Conference – ANTEC, Orlando, FL, USA, 2000, Paper No.578.

9.

H-C. Kuan, J-M. Huang, C-C.M. Ma and F-Y. Wang, Plastics Rubber and Composites, 2003, 32, 3, 122.

10. E. Kowalska, Progress in Rubber Plastics and Recycling Technology, 2002, 18, 3, 173.

271

Mixing in Single Screw Extrusion

272

A

ppendix – Preparation of Microtome Sections for Assessment of Dispersive and Distributive Mixing

The instrument used to prepare microtomed sections reproduced as photmicrographs in Chapters 2, 3, 7, 8 and 9 was a Slee (Reickert) sledge microtome. Although disposable plate glass and sharpenable steel blades are very good in the hands of experts, time constraints dictated that disposable blades were used. Whichever is used, great care must be taken with regards to both personal safety and prevention of damage to the very sharp cutting edge. The technique used was that described by James in Microscopy Notes 6, which is reproduced in full here.

Microscopy Notes 6 - Ivan James, Rapra Bulletin, 1971, 25, 4, p.85-87 (Reproduced with permission from Smithers Rapra, Shawbury, Shrewsbury, UK. ©1971, Rapra Technology)

Flattening Sections Satisfactory sections of thermoplastics may have any thickness between 3 µm and 20 µm depending on the material. They can be cut relatively easily but are rarely flat enough for immediate mounting. Usually they form tight rolls which resist all attempts to brush them flat, or occasionally, with very thin sections, they may concertina instead. Lack of flatness renders examination difficult, and the corresponding change of focus from one part of the field to another means that any photographs taken are of poor quality. Biologists cutting paraffin wax sections are not faced with quite the same problem. Paraffin wax remains fairly flat during cutting and sections adhere one to another to form a long ribbon. Corrugations in the sections are removed by floating them on the surface of water which is warm enough for the surface tension forces to stretch the sections until they are flat. For this technique to be successfully applied to plastics it is necessary to replace the water with a liquid which boils above the softening points of most plastics and which has a high surface tension at these temperatures. Add the further requirement that the liquid must not dissolve or attack plastics and the search seems daunting. Happily, glycerol fulfils all the requirements and has the added advantages of total miscibility 273

Mixing in Single Screw Extrusion with water and a density sufficiently high for many plastics to float in it without the aid of surface tension. Physical Properties of Glycerol Density: 1.26 g/cm3 at 20 °C Boiling Point: 290 °C Surface Tension at 20 °C: 6.34 Pa; at 150 °C: 5.19 Pa Glycerol is totally miscible with water and alcohol. Sections of any thermoplastic material can be unrolled on glycerol and the technique is also successful with some thermosets. Details of the method are set out next.

Trimming the Block Although sections of soft thermoplastics such as polyethylene cut easily, they do not separate cleanly from the block but hinge back on a thin strip of material at the back edge (Figure 1). This can be overcome by trimming the rear of the block to a point so that there is then insufficient area for a hinge to form (Figure 2).

Figure 1 Section at rear of the block.

Figure 2 Rear of the block trimmed to a point. 274

Preparation of Microtome Sections for Assessment of Dispersive and Distributive Mixing

Figure 3 Drop of glycerol forming a seal

Figure 4 Block trimmed to a ‘tie’ shape.

Figure 5 Rolled section on a pool of glycerol. Contact zone.

Another difficulty which can arise when a rolled section is placed on a pool of glycerol is that a drop of liquid bridges the gap between the first two turns (Figure 3). If this happens the section will not subsequently unroll at all but will instead collapse as a thick sandwich. This can be avoided by trimming the forward part of the block as well, so that it has a ‘tie’ shape (Figure 4). The rolled part of the section can then be kept within the outline formed by the contact of the section with the glycerol (Figure 5). This is by far the easiest section to deal with and it is recommended that blocks should be trimmed to this shape whenever possible. Dimensions are not critical, but a width of 4 mm and a length of 12 mm is probably adequate in most cases. 275

Mixing in Single Screw Extrusion

Flattening the Rolled Sections Sections should be cut dry, and then picked up with a sable brush and put down carefully so that the broadest part of the roll just touches the surface of a small pool of glycerol on a microscope slide. If, now, the microscope slide be warmed, either on a hot plate set at a low temperature (say 150 °C) or above a low Bunsen flame, then as the thermoplastic softens the surface tension forces will pull the section down into the surface and render it completely flat. Rarely can a problem have been so efficiently solved by the simple application of physical principles.

Holey Sections Sometimes it happens that in spite of the care taken, sections from a particular block will not flatten, but always collapse in a sandwich. Almost invariably it will be found that the section contains one or two small holes through which glycerol seeps, thus glueing the layers of the roll together. Since the holes may indicate porosity and will almost certainly be relevant to the investigation in hand, it is pointless seeking an area free from them and another method of flattening must be used.

Brushing Flat For this method a closer control of temperature is needed and it is advisable to use a hot plate rather than a Bunsen flame. The low temperatures involved are most easily achieved by wiring the hot plate to a ‘Reguplug’ simerstat control. A section is placed in a shallow dish containing a small quantity of cold glycerol. The dish is then put onto the hot plate, the temperature of which should not initially exceed 100 °C. As the section warms an attempt should be made to unroll it using two sable brushes. Initially the section will spring back into a roll, but if the temperature is increased very slowly a point will be found where it can be brushed flat without sticking either to itself or to the brushes. The range of temperature over which this operation can be carried out is quite small and it is advisable not to attempt to heat the glycerol too quickly. Once the section is nominally flat it can be transferred onto a pool of glycerol on a microscope slide and relaxed completely, if desired.

Distortion The distortion of sections during cutting is familiar to all microtomists and is commonly referred to as ‘compression’. In the case of a sample which is initially free from internal stresses warming on glycerol will completely remove the compression due to cutting and will restore the section to its original shape and size. If, as is often the 276

Preparation of Microtome Sections for Assessment of Dispersive and Distributive Mixing case, the block has some frozen in stresses, then not only will the compression due to cutting be removed but the frozen in stresses also, with the result that the final section may not be the same shape as the original block. In most cases this does not matter, but in those cases where it is important the brushing technique should be used.

Washing and Mounting Glycerol is a poor mountant and is best removed from the sections, but no attempt should be made to do this while the glycerol is hot. Once the slide has cooled down it should be dipped into a trough of water. The plastic sections will float on the water surface and the glycerol will disperse: the disappearance of the streamers which can be seen in the water is an adequate indication that the section is free of glycerol. If a large number of sections is being examined then transfer to a second bath of water is desirable. At this stage the sections are quite robust and can be dried and mounted in a conventional medium without difficulty. It is probably worth pointing out, however, that many of the media used for biological mounting (e.g., clove oil) will swell or dissolve certain thermoplastics. Thus chemical activity rather than refractive index may be the deciding factor in the selection of a mountant. An interesting example of rotationally cast black polyethylene by this method is shown in Figure 6.

Figure 6 Rotationally cast polyethylene showing incomplete mixing of black and natural phases. 277

Mixing in Single Screw Extrusion

278

A

bbreviations

ABS

Acrylonitrile-butadene-styrene

BS

British standards

CA

Cellulose acetate

CBA

Chemical blowing agent(s)

CFC

Chlorofluorocarbon(s)

CRD

Chris Rauwendaal mixer

CTM

Cavity transfer mixer(s)

D

The distance along an extruder screw in terms of screw diameters

EFM

Elongational flow mixer

EPDM

Ethylene-propylene diene terpolymer

ET

Energy transfer

EVOH

Ethylene vinyl alcohol

h

Height

HDPE

High-density polyethylene

HFC

Hydrofluorocarbon(s)

HIPS

High impact polystyrene

HMWPE

High molecular weight polyethylene

HTPC

High temperature polycarbonate

L/D

Length to diameter ratio

L/H

Length to height ratio

LDPE

Low-density polyethylene

LLDPE

Linear low-density polyethylene

MDPE

Medium density polyethylene

OD

Outer diameter

PA

Polyamide(s)

PA11

Polyamide-11

PA12

Polyamide-12 279

Mixing in Single Screw Extrusion PA6

Polyamide-6

PBT

Polybutylterephthalate

PC

Polycarbonate

PE

Polyethylene(s)

PEEK

Polyether ether ketone

PET

Polyethylene terephthalate

PP

Polypropylene

PPO

Polyphenylene oxide

PPS

Polyphenylene sulfide

PPVC

Plasticised polyvinyl chloride

PS

Polystyrene

PVC

Polyvinylchloride

rpm

Revolutions per minute

RTD

Residence time distribution

SAN

Styrene-acrylonitrile

SEM

Scanning electron microscopy

SMA

Styrene-maleic anhydride

STP

Standard temperature and pressure

TMR

Twente mixing ring

UL

Underwriters’ Laboratory

UV

Ultraviolet

VHMWHPDE Very high molecular weight high-density polyethylene VHMWPE

280

Very high molecular weight polyethylene

I

ndex

A A2-B2 mixer 44, 46 Agglomerate count 62 Agglomerate dispersion 24 Agglomerate formation 232, 234 Agglomerate measurement 61 Agglomerates 14, 17, 19, 20, 23, 24, 26, 54, 61-63, 68, 208, 211, 229, 230, 233, 235, 238, 239, 247, 250, 261, 263

B Barmag key slot mixer 173 Barmag mixers 184 Barr energy transfer screws 126 Barr floating ring mixer 195 Barr multi-ring mixer 195 Barr ring mixer 194 Barrier flight melting screws 115 Barrier flight screws 127 Barrier screw 19, 73, 75, 107, 116, 121-125, 128, 129, 233 North American 118 Blending 20 Blending, polymer 243, 246, 250, 254, 255, 258, 261 Brushing flat 276

C Compression melting zone 103 Couette flow 40, 41, 243, 249 Crushing tests 27

281

Mixing in Single Screw Extrusion

D Dispersion, by brittle fracture 256 Dispersion theory 247 Double wave screw 125

E Electron microscopy 69 Erwin’s mixing model 43 Mixing, extensive 20 Extrusion foaming 220, 221, 223

F Feed conveying 91 Feed zone 87, 91, 92, 93, 94, 95, 96, 99, 100, 103, 129, 130, 205, 206 Feed zone screw cooling 99 Fibre extrusion 213 Fibrils 252, 254, 256 Mixer, floating ring 189, 194, 196 Friction tester 97

G Gels 14, 15, 31, 68, 147, 261, 263 Gels, elimination of 262 Gerber mixers 171 Gravity conveying 234 Grooved feed zone 90, 123, 128, 129

H Haake rheocord torque rheometer 250 Holey sections 276 Honeycomb packing 197 Hoppers, design 78 Hoppers, flow test 84, 85 Hoppers, round 80, 81 Hoppers, square 81

282

Index

I Image analysis 61, 62 Immiscible blends 244 Liquid additives, incorporation of 204 Infrared spectroscopy 68 Injection moulding 5, 16, 189, 191, 193, 244, 270 Check ring 190 Check ring cavity mixer 192 Check ring mixers 189

K Kenics mixer 198, 199

L Lamina flow mixing 34 Laminar shear 39, 75 Flow 247, 248 Flow mixing 29 Flow models 243 Mixing 32, 35, 54, 75, 176, 254 Liquid injection processes 208

M Maddock element screw 108 Maillefer barrier screw 117, 124 Mass flow hopper 78 Meillefer barrier screw 118, 119 Melt conveying 120 Melt filled screw channel 137 Melt flow index tester 31 Melt pumping zone 103 Melting mechanism 101, 102, 103, 107 Metal box key slot mixer 172 Miscible blends 244 Mixer, cavity transfer 19, 52, 53, 168, 191, 214, 215, 233, 220, 256 Model 49 Rotor 224

283

Mixing in Single Screw Extrusion Mixer, check ring 193 Mixer, Couette type 256 Mixer, elongational flow 269 Mixer, helical 197, 199 Mixer, honeycomb 198, 199 Mixer, interacting 167 Mixer, Maddock 148, 196, 233 Mixer, pin 170 Mixer, pineapple 160, 255 Mixer, planetary gear 233 Mixer, radial static dispersive 262 Mixer, Renk (barmag) 172 Mixer, Ross ISG 198 Mixer, rounded cavity 176 Mixer, Stanley 168 Mixer, Stanley turbine 42 Mixer, static 197, 200, 201 Mixing, degree of 33 Mixing, dispersive 5, 15, 17, 18, 20, 21, 24, 25, 29, 54, 59, 161, 229, 230, 239, 243, 246, 269, 273 Mixing, distributive 5, 6, 8, 11, 13, 15, 17, 18, 20, 23, 29, 32, 33, 37, 59, 67, 125, 135, 161, 167, 176, 197, 262, 263, 273 Mixing effects 71 Mixing, elongational 20 Mixing, elongational flow 261 Mixing, intensive 20 Mixing, Maddock 189 Mixing, masterbatch 11 Mixing, measurement 59 Mixing, mechanism 197 Mixing, pin 189 Mixing, pins 149, 150 Mixing, screw channel 135 Mixing, simple 20

N Non-mass flow hopper 79, 82 Nut-on-bolt model 87

284

Index

O Optical microscopy 69

P Particulate friction measurements 96 Pellet handling 77 Plastics industry 128 Plug conveying 89 Plug flow conveying 234 Polymer industry 7 Polymer scrap 246 Polymer waste 246, 259 Polymerisation 244 Poppet valve liquid injector 206 Pumping zone 105

R Rectangular hoppers 81 Reifenhauser staromix 184, 185, 186 Residence time distribution 144, 146, 147 Retarded taylor dispersion 255 Rubber industry 171

S Scanning electron microscopy 250, 251 Screen pack extrusion test 63 Screen pack filtration test 62 Screw design 103, 131 Screw, melting 116 Shear heating 177 Shear-ring screw 127 Silage wrap film 14, 15, 203, 208 Silane grafting 218, 219 Silicone lubricant injection 220

285

Mixing in Single Screw Extrusion Single screw extruder 8, 9, 12, 17, 19, 20, 24, 25, 29, 35, 38, 54, 59, 61, 135, 147, 210, 230, 240, 243, 250-252, 254, 256, 259, 261, 262, 269, 270 Mixing 23 Stages 71 Single screw extrusion 5, 60, 254, 257, 258 Solids bed break-up 105, 106, 107, 117 Solids conveying 87, 88, 89, 99, 121, 234 Starved feeding 234, 237, 239 Stratablend mixing screw 126, 127 Striation formation 135 Striation thickness measurement 60, 182

T Taylor-like droplet dispersion 255 The scott and macosco mechanism 256 Thermoforming 243, 244 Three pump spectrum system 216 Torque rheometer 251 Turbine mixing heads 168, 169 Turbine system 43 Twente mixing ring 161 Twin screw 259 Twin screw extruder 8, 12, 28, 153, 223, 261, 270 Twin screw extrusion 262

V Variable barrier energy transfer screw 126

W Woodroffe key slot mixers 171

286

Published by Smithers Rapra, 2009

A lot of plastics products are extruded and most of those products contain additives, which mean that the plastic must be mixed at some point. Mixing is generally done with a twin-screw extruder, and the single-screw extruder, which is used in product production, is generally overlooked as a device for mixing. This reference handbook, written by a former Principal Consultant at Smithers Rapra, and the inventor of the cavity transfer mixer, redresses the balance. This handbook is a must-have practical guide to the subject of single-screw extrusion. Avoiding mathematical theory, except when absolutely necessary, this authoritative handbook empowers the reader to achieving good results with their plastic mixing. Mixing in Single Screw Extrusion will be a valuable resource to all involved in the art of plastic extrusion.

Shawbury, Shrewsbury, Shropshire, SY4 4NR, UK Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118 Web: www.ismithers.net