Advances in Urethane Science and Technology

Citation preview

Advances in Urethane Science and Technology Editors: D. Klempner K.C. Frisch

Advances in Urethane Science and Technology

Daniel Klempner and Kurt Frisch

Rapra Technology Limited Shawbury, Shrewsbury, Shropshire, SY4 4NR, United Kingdom Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118 http://www.rapra.net

First Published in 2001 by

Rapra Technology Limited Shawbury, Shrewsbury, Shropshire, SY4 4NR, UK

©2001, Rapra Technology Limited

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.

ISBN: 1-85957-275-8

Typeset by Rapra Technology Limited Printed and bound by Lightning Source UK

Dedication

In memory of Kurt C Frisch One of the founding Fathers of polyurethanes

January 15th 1919 to October 21st 2000

Contents

1 Dimensional Stabilising Additives for Flexible Polyurethane Foams ................ 3 1.1 Introduction ............................................................................................. 3 1.2 Experimental Procedures .......................................................................... 6 1.2.1 Materials ....................................................................................... 6 1.2.2 Handmix Evaluations .................................................................... 8 1.2.3 Machine Evaluation ...................................................................... 9 1.3 TDI - Flexible Moulded Additives .......................................................... 15 1.3.1 Dimensional Stability Additives for TDI ...................................... 16 1.3.2 Low Emission Dimensional Stability Additives ............................ 42 1.4 MDI Flexible Moulded Foam Additives ................................................. 63 1.4.1 Dimensional Stability Additives for MDI .................................... 64 1.4.2 Low Emissions Dimensional Stability Additives in MDI.............. 67 1.5 TDI Flexible Slabstock Low Emission Additives ..................................... 73 1.5.1 Reactivity .................................................................................... 74 1.5.2 Standard Physical Properties ........................................................ 74 1.5.3 TDI Flexible Slabstock Foam Review .......................................... 74 1.6 Foam Model Tool Discussions ................................................................ 75 1.6.1 TDI and MDI Moulded Foam Model .......................................... 75 1.6.2 TDI Flexible Slabstock Foam Model ........................................... 78 1.7 Conclusions ............................................................................................ 81 2 Demands on Surfactants in Polyurethane Foam Production with Liquid Carbon Dioxide Blowing .................................................................... 85 2.1 History of Polyurethane Foams .............................................................. 85 2.1.1 Environmental Concerns in Relation to Flexible Foam Density ... 86

i

Advances in Urethane Science and Technology

2.2 Current Liquid Carbon Dioxide Technologies for Flexible Slabstock ......... Polyether Foam Production .................................................................... 88 2.2.1 Machinery ................................................................................... 88 2.2.2 The Foaming Process ................................................................... 90 2.2.3 Additional Tasks of Silicone Surfactants in Flexible Slabstock Foam Production ......................................................................... 95 2.2.4 Chemistry of a Silicone Surfactant in Flexible Slabstock Foam Production ......................................................................... 99 2.2.5 A Surfactant Development Example .......................................... 101 3 Polyurethane Processing: Recent Developments ........................................... 113 3.1 Industrial Solutions for the Production of Automotive Seats Using Polyurethane Multi-Component Formulations ........................... 113 3.1.1 Market Requirements ................................................................ 113 3.1.2 Dedicated Solutions: Metering Equipment ................................ 114 3.1.3 Dedicated Solutions: Mixing Heads........................................... 116 3.1.4 Dedicated Solutions ................................................................... 121 3.2 ‘Foam & Film’ Technology - An Innovative Solution to Fully Automate the Manufacture of Automotive Sound Deadening Parts ..... 130 3.2.1 The Problem .............................................................................. 131 3.2.2 The Approach to a Solution ...................................................... 131 3.2.3 The Film .................................................................................... 133 3.2.4 Industrial Applications .............................................................. 135 3.2.5 Applications .............................................................................. 137 3.2.6 Advantages ................................................................................ 138 3.3 InterWet - Polyurethane Co-injection ................................................... 138 3.3.1 Glass-Reinforced Polyurethanes, a Well-Known Technology ..... 139 4 Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels for Super Insulation Applications ...................................................... 157 4.1 Introduction ......................................................................................... 157 4.2 Some General Properties of Open Cell PU Foams for Vacuum

ii

Contents

Insulated Panels .................................................................................... 158 4.3 Vacuum Issues in the Selection of VIP Components .............................. 163 4.3.1 Vacuum Properties of the Open Cell Foams .............................. 163 4.3.2 Vacuum Properties of the Barrier Film ....................................... 167 4.3.3 The Getter Device ...................................................................... 179 4.4 Vacuum Panel Manufacturing Process and Characterisation ................ 188 4.4.1 Some Manufacturing Issues ....................................................... 188 4.4.2 Characterisation of Vacuum Panels ........................................... 191 4.5 Insulation Performances of Open Cell PU-Filled Vacuum Panels .......... 196 4.6 Examples of VIP Applications and Related Issues................................. 199 4.6.1 Household Appliances ............................................................... 199 4.6.2 Laboratory and Biomedical Refrigerators .................................. 203 4.6.3 Vending Machines ..................................................................... 204 4.6.4 Refrigerated/Insulated Transportation ....................................... 205 4.6.5 Other Applications .................................................................... 206 4.7 Near Term Perspectives and Conclusions.............................................. 206 5 Modelling the Stabilising Behaviour of Silicone Surfactants During the Processing of Polyurethane Foam: The Use of Thin Liquid Films ................. 213 5.1 Introduction ......................................................................................... 213 5.2 Film Drainage Rate: Reynold’s Model and Further Modifications ........ 216 5.2.1 Rigid Film Surfaces .................................................................... 216 5.2.2 Mobile Film Surfaces ................................................................. 217 5.2.3 Surface Viscosity ........................................................................ 217 5.2.4 Surface Tension Gradients ......................................................... 218 5.3 Experimental Investigation of Model, Thin Liquid Polyurethane Films and the Development of Qualitative and Semi-Quantitative Models of Film Drainage ...................................................................... 219 5.3.1 Experimental Details ................................................................. 221 5.3.2 Qualitative Description of Polyurethane Films .......................... 223

iii

Advances in Urethane Science and Technology

5.3.3 Quantitative Measurement of Film Drainage Rates: Bulk and Surface Effects ............................................................ 226 5.4 The Development of Theoretical Models of Vertical, Draining Thin Liquid Model PU Films ................................................................ 236 5.4.1 Rigid-Surfaced Collapsing Wedge Model ................................... 236 5.4.2 Deforming Film Models ............................................................ 239 5.4.3 Tangentially-Immobile Films ..................................................... 242 5.4.4 Finite Surface Viscosity .............................................................. 245 5.4.5 Adding Surfactant Transport ..................................................... 249 5.5 Summary .............................................................................................. 254 6 Synthesis and Characterisation of Aqueous Hybrid Polyurethane-UreaAcrylic/Styrene Polymer Dispersions ............................................................ 261 6.1 Preface .................................................................................................. 261 6.2 Introduction ......................................................................................... 261 6.2.1 General Considerations ............................................................. 261 6.2.2 Acrylic Dispersions and Polyurethane Dispersions (DPUR) ....... 264 6.2.3 Hybrid Acrylic-Urethane Dispersions ........................................ 266 6.3 Concept of the Study ............................................................................ 268 6.3.1 Selection of Starting Materials ................................................... 268 6.3.2 Assumptions for Synthesis of Hybrid Dispersions ..................... 269 6.4 Methods of Testing ............................................................................... 276 6.4.1 Dispersions ................................................................................ 276 6.4.2 Coatings .................................................................................... 277 6.4.3 Films .......................................................................................... 278 6.5 Experimental results ............................................................................. 279 6.5.1 Characterisation of Starting Dispersions Used for Synthesis of MDPUR ................................................................................ 279 6.5.2 Synthesis of MDPUR and MDPUR-ASD ................................... 288 6.5.3 Investigation of the Effect of Various Factors on the Properties of Hybrid Dispersions ............................................... 290

iv

Contents

6.5.4 Additional Experiments ............................................................. 312 6.6 Discussion of results ............................................................................. 320 6.6.1 Estimation of the Effect of Various Factors on the Properties of Hybrid Dispersions and Films and Coatings Made from Them ................................................................................ 320 6.6.2 Mechanism of Hybrid Particle Formation ................................. 326 6.7 Summary ................................................................................................ 330 7 Adhesion Behaviour of Urethanes ................................................................ 335 7.1 Introduction ......................................................................................... 335 7.2 Surface Characteristics of PU Adhesive Formulations ........................... 335 7.2.1 Experimental ............................................................................. 336 7.2.2 Results and Discussion .............................................................. 338 7.2.3 Conclusions ............................................................................... 347 7.3 Acid/Base Interactions and the Adhesion of PUs to Polymer Substrates 347 7.3.1 Experimental ............................................................................. 348 7.3.2 Results and Discussion .............................................................. 351 7.4 The Effectiveness of Silane Adhesion Promoters in the Performance of PU Adhesives .................................................................................... 355 7.4.1 Experimental ............................................................................. 356 7.4.2 Results and Discussion .............................................................. 358 7.4.3 Conclusions ............................................................................... 364 8 HER Materials for Polyurethane Applications ............................................. 369 8.1 Introduction ......................................................................................... 369 8.2 Experimental Conditions ...................................................................... 370 8.2.1 Chain Extenders ........................................................................ 370 8.2.2 Prepolymers ............................................................................... 370 8.2.3 Preparation of Cast Elastomers ................................................. 372 8.2.4 Physical and Mechanical Properties Determination ................... 372 8.3 HER Materials Synthesis and Characterisation .................................... 373 v

Advances in Urethane Science and Technology

8.4 Cast Poly(Ether Urethanes) ................................................................... 375 8.4.1 Pot Life Determination .............................................................. 375 8.4.2 Polyurethane Castings ............................................................... 376 8.4.3 Calculation of Hard and Soft Segment Contents ....................... 376 8.4.4 Hard Segment Versus Hardness ................................................. 378 8.4.5 Tensile Properties ....................................................................... 378 8.4.6 Tear, Compression Set and Rebound Properties......................... 380 8.4.7 Differential Scanning Calorimetry ............................................. 381 8.4.8 Dynamic Mechanical Analysis ................................................... 383 8.5 Cast Poly(Ester Urethanes) ................................................................... 390 8.5.1 Pot Life Determination .............................................................. 390 8.5.2 Tensile Properties ....................................................................... 390 8.5.3 Tear, Fracture Energy, Compression Set and Rebound Properties ................................................................... 390 8.5.4 Differential Scanning Calorimetry Analysis ............................... 393 8.5.5 Dynamic Mechanical Analysis ................................................... 395 8.6 Cast Polyurethanes from HER/HQEE Blends ....................................... 397 8.6.1 Freezing Point Determination of HER/HQEE Blends ................ 397 8.6.2 Cast elastomers and Their Properties......................................... 398 8.7 High Hardness Cast Polyurethanes ....................................................... 401 8.7.1 Cast Elastomers and Their Hard and Soft Segment Contents .... 401 8.7.2 Hardness, Tensile, Tear, Compression Set and Rebound Properties ................................................................... 401 8.7.3 FT-IR Analysis of Cast Polyurethanes ........................................ 403 8.7.4 Differential Scanning Calorimetric Analysis .............................. 405 8.7.5 Dynamic Mechanical Analysis ................................................... 405 8.8 High Thermal Stability Polyurethane with Low Heat Generation ........ 405 8.8.1 Hardness Measurements ............................................................ 408 8.8.2 Tensile Measurements ................................................................ 408 8.8.3 Differential Scanning Calorimetric Analysis .............................. 410

vi

Contents

8.8.4 Dynamic Mechanical Analysis ................................................... 412 8.9 Conclusions .......................................................................................... 416 9 Ultra-Low Monol PPG: High-Performance Polyether Polyols for Polyurethanes ......................................................................................... 421 9.1 Introduction ........................................................................................... 421 9.2 MDI/BDO Cured Elastomers Based on Ultra-Low Monol PPG Polyols . 424 9.2.1 Effect of Monol Content on 4,4´-Methylene Diphenylmethane Diisocyanate (MDI)/1,4-Butanediol (BDO) Cured Elastomers... 424 9.2.2 Processability and Property Latitude of Elastomers Based on Ultra-Low Monol PPG Polyols .................................................. 429 9.2.3 Processing Latitude Improves by Incorporating Oxyethylene Moieties ................................................................ 434 9.3 One-Shot Elastomer System Based on EO-Capped, Ultra-Low Monol PPG Polyols .............................................................................. 436 9.3.1 Effect of Primary Hydroxyl Concentration on One-Shot Elastomer Processability ............................................................ 436 9.3.2 Effect of Monol Content on One-Shot Elastomer Processability and Properties............................................................................ 438 9.3.3 Processability and Property Latitude of Elastomers Based on EO-Capped, Ultra-Low Monol Polyols ................................ 440 9.4.1 MDI/BDO Cured Elastomers: Acclaim Polyol 3205 Versus ............. PTMEG-2000 ............................................................................ 445 9.4.2 Enhanced Elastomer Properties Utilising Ultra-Low Monol PPG/PTMEG Blends .................................................................. 447 9.5 Polyol Molecular Weight Distribution Effect on Mechanical and Dynamic Properties of Polyurethanes ............................................ 449 9.5.1 TDI Prepolymers Cured with Methylene Bis-(2-Chloroaniline) [MBOCA] .................................................................................. 450 9.5.2 Moisture-Cured TDI Prepolymers ............................................. 454 9.5.3 Aqueous Polyurethane/Urea Dispersion Coatings ...................... 456 9.5.4 MDI Prepolymers Cured with BDO .......................................... 459 9.6 Conclusions .......................................................................................... 461

vii

Advances in Urethane Science and Technology

APPENDIX........................................................................................................ 465 Laboratory Preparation of 2,4-TDI and 4,4´-MDI Prepolymers ................... 465 Laboratory Casting of 4,4´-MDI Prepolymers Cured with BDO .................. 465 Laboratory Casting of One-Shot Elastomers Based on CarbodiimideModified MDI, Polyol, and BDO ................................................................. 465 Laboratory Casting of 2,4-TDI Prepolymers Cured with MBOCA .............. 466 Laboratory Moisture-Curing of 2,4-TDI Prepolymers ................................. 466 Laboratory Preparation of Aqueous Polyurethane/Urea Dispersions using the Prepolymer Mixing Process ........................................................... 466 Abbreviations .................................................................................................... 469 Contributors ...................................................................................................... 473 Author Index ..................................................................................................... 477 Main Index ........................................................................................................ 483

viii

Preface

This is a landmark issue of ‘Advances in Urethane Science and Technology’. Not only is this the first volume of the new millennium, but it is the first to be published by Rapra Technology. On a more solemn note, one of the editors, Kurt C. Frisch, passed away shortly before publication. Dr. Frisch, founder of the University of Detroit Mercy’s Polymer Institute, was one of the pioneers of polyurethanes and was responsible for the successful introduction of polyether polyurethane flexible foams into commerce in the mid-1950s. Let us not only mourn the loss of, but also celebrate the life of this great scholar by continuing to further the frontiers of urethane science and technology. This volume is a good example of this progress. Polyurethanes continue to be one of the most versatile of all polymers, finding applications in foams (flexible, rigid, and in-between), elastomers, coatings, sealants, adhesives, paints, textiles, and films. This volume presents some of the major advances in polyurethanes, both from the materials and research side of things as well as processing and applications, and includes studies on foams (additives, vacuum panel applications, blowing and processing), elastomers, adhesion behaviour and new urethane raw materials. I would like to take this opportunity to express my gratitude to the authors who contributed to this book and to the University of Detroit Mercy for its encouragement of this effort. I would also like to thank the staff of Rapra, in particular, Frances Powers, Claire Griffiths and Steve Barnfield. Daniel Klempner, Ph. D. Polymer Institute,University of Detroit Mercy July, 2001

1

Advances in Urethane Science and Technology

2

1

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Gary D. Andrew, Jane G. Kniss, Mark L. Listemann, Lisa A. Mercando, James D. Tobias and Stephan Wendel

1.1 Introduction The issues that an automotive seat manufacturer faces when formulating and producing seats are escalating. Physical properties such as tensile and tear strengths, compression set and wet set are critical when meeting specific mechanical performance requirements as defined by the original equipment manufacturer (OEM). As new requirements for comfort and durability are instituted, tests such as dynamic creep testing, long term vibration characterisation and repeated compression tests under various atmospheric and load conditions have been used to characterise foam performance for comfort. Comfort properties are best controlled by the polyols used to produce the polyurethane foam cushion. Significant changes in polyol technology to meet these dynamic comfort properties have had an impact on the processing of polyurethane foam and on physical properties. Increased tightness of the foam article resulting from changes in these raw materials has focused more attention by foam producers on crushing methods. Flexible moulded polyurethane foam requires some type of mechanical crushing to prevent shrinkage and ultimately maintain part stability. With recent changes made to polyol technology, mechanical methods of crushing do not always provide the consistency required to produce a part that is dimensionally stable. Additionally, producers of polyurethane articles are continually building more complexity into their seat designs to meet the aesthetic values required by today’s consumers. These complex seat designs place more emphasis on crushing capability due to the nature of the designs. With all these changes, additives needed to be developed which provide a wider processing latitude and increased breathability to the polyurethane article. Wider processing latitudes should reduce scrap and repair rates on the foam production line and improve economics for the polyurethane producer [1]. The formation of moulded foam is a complicated chemical process which involves several reactions occurring simultaneously. There are rapid volume and temperature increases and the concurrent development of phase separated polymer networks. To understand how foam properties can be affected by catalyst and surfactant chemistries several techniques are used to identify key performance benefits and issues. A force to crush

3

Advances in Urethane Science and Technology (FTC) detection device was used to measure the force required to crush a part to 50% of its thickness for determination of cell openness. Mass-loss/rate-of-rise was run to understand rate of rise and height measurements, weight loss from carbon dioxide generation and temperature profiles. A scanning electron microscope (SEM) was used to determine differences in cell structure and cell distribution caused by changes in the catalyst and surfactant chemistries. Physical properties were also tested using ASTM test methods for flexible cellular polyurethane. A novel chemical reaction foam modelling technique was also used to determine the selectivity of the catalyst packages, compared to industrial standard controls [2]. In the past it was thought that the cell structure of polyurethane foam is controlled by the type and amount of surfactant used. Dabco DC5043 (Air Products and Chemicals, Inc.) was developed to enhance cell wall drainage to better enable cell opening during crushing cycles. It was also thought that surfactant technology was the best way to provide improved crushing techniques; therefore, catalyst technology was ignored [3]. As mentioned earlier, with new polyol technology development more emphasis was placed on crushing. New additive technology needed to be developed that would open cells during the foam formation and reduce the requirement and criticality of the crushing processes. The technology had to go beyond providing easier cell opening at crush to providing more open cells during the polyurethane formation. The real challenge in polyurethane foam formation is to control the chemical and physiochemical processes up to the point where the material finally sets. The sequence and the rate of the chemical reactions are predominately a function of the catalyst and the reactivity of the basic raw materials, polyol and isocyanate. The physiochemical contribution to the overall stability and processability of a system is provided by the silicone surfactants. Optimum foaming results will be achieved only if the correct relationship between chemistry and physics exists [4]. Another rapidly increasing environmental concern is over the emission of volatile organic compounds (VOC) during and following the production of industrial and consumer goods. This has stimulated a great deal of effort within the chemical industry to reduce and/or control the ways in which such emissions may occur. In the polyurethane foam industry, efforts to reduce VOC emissions have greatly impacted the technologies used in manufacturing processes, especially for the use of organic auxilliary blowing agents such as chlorofluorocarbons. In addition, the ultimate fate of additional foaming additives, including surfactants and catalysts, is now coming under increased global scrutiny. As a result, foam manufacturers have expressed a desire for polyurethane additives that, among other things, do not exhibit the degrees of fugitivity common to many of the additives that are currently used in polyurethane foam production today.

4

Dimensional Stabilising Additives for Flexible Polyurethane Foams Polyurethane foams are prepared from the simultaneous reactions of diisocyanate with water and with polymeric diols and/or triols to form hydrogen-bonded urea (hard) segments and polyurethane networks (soft segments). The commercial production of polyurethanes via isocyanate poly-addition reactions requires the use of one or more catalysts. Tertiary amines are widely accepted in the industry as versatile polyurethane catalysts. Amine catalysts are generally stable in the presence of standard polyurethane formulation components and can have an impact on both the blowing (water-isocyanate) and gelling (polyol-isocyanate) reactions. Although the use of catalysts in the manufacturing of polyurethane foam both speeds the production of the foamed article and, through the judicious choice of catalyst package, allows control of the physical properties of the product, there are some problems associated with the use of these additives. A number of commonly used tertiary amine catalysts can volatilise under certain conditions. Release of tertiary amines during foam processing and from consumer products is generally undesirable. Therefore, identifying alternatives to standard tertiary amine catalysts which have no or low volatility, yet exhibit the same type of activity in isocyanate poly-addition reactions, is desirable. The non-fugitive catalysts reported in this chapter address the problems associated with the use of polyurethane catalysts by reducing the odour and volatility of these materials and by eliminating the ability of these additives to escape from finished foam products. One strategy has involved functionalising the catalysts to render the species reactive toward isocyanates, thereby covalently attaching the catalysts to the polymer network. This strategy not only renders the catalytic material non-fugitive in the final product, but also reduces the odour and volatility of the catalyst through increases in molecular weight and polarity. These non-fugitive catalysts also provide equivalent or improved physical properties when compared to industry standards, whereas conventional reactive amine catalysts as well as metal catalysts cannot always meet todays ever increasing manufacturer and consumer performance requirements. These increasingly evolving requirements have led to the development of both novel non-fugitive catalysts and new cell-opening non-fugitive catalysts for flexible foam. These new low emission additives have been developed to meet the challenge of optimised foaming and result in little or no emissions. Several of the non-fugitive catalysts possess cell-opening capability. This new technology allows the manufacturer of polyurethane foam to optimise their system to achieve the best processing latitude for their foam process. These new additives maintain, or in some cases, improve key physical properties while providing a more open foam.

5

Advances in Urethane Science and Technology

1.2 Experimental Procedures Data presented herein was derived from a combination of handmix and high pressure impingement-mixing machine produced foam. Foams were prepared using several general types of formulations for toluene diisocyanate (TDI) and two general types of formulations for methylenediphenyl diisocyanate (MDI) which are representative of currently utilised formulations in the automotive interior component industry. In addition, an all water blown formulation was used to represent the flexible slabstock industry.

1.2.1 Materials The materials used are shown in Table 1.1.

Table 1 Materials used in experimental work Trade name

Formulation

Manufacturer

Dabco 33LV

33% TEDA in DPG

APCI

Dabco BL-17

Delayed action tertiary amine blowing catalyst.

APCI

Dabco BL-53

Newly developed tertiary amine, which provides blowing and cell opening capabilities.

APCI

Dabco B-16

Tertiary amine surface cure catalyst.

APCI

Polycat 15 (PC-15) Balanced reactive amine catalyst.

APCI

Dabco BLV

Dabco 33-LV/Dabco BL-11 in a 3:1 catalyst blend.

APCI

Dabco T-9

Stannous octoate catalyst.

APCI

Dabco NE1060

Newly developed non-fugitive gelling catalyst for APCI flexible moulded applications.

XF-N1085

Newly developed non-fugitive cell opening blowing catalyst.

Dabco NE500

Newly developed non-fugitive gelling catalyst for APCI flexible slabstock foam from APCI.

Dabco NE600

Newly developed non-fugitive intermediate blowing catalyst for flexible slabstock applications.

APCI

Dabco NE200

Newly developed non-fugitive intermediate blowing catalyst for flexible slabstock applications.

APCI

6

APCI

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1 Continued Trade name

Formulation

Manufacturer

XF-O11006

Newly developed non-fugitive cell opening gelling catalyst.

APCI

Dabco DC5169 Silicone copolymer surfactant for cold cure systems.

APCI

Dabco DC5043 Silicone copolymer surfactant for TDI cold cure systems.

APCI

Dabco DC2585 Silicone copolymer surfactant for MDI cold cure systems.

APCI

Dabco DC2517 Silicone copolymer surfactant.

APCI

Dabco DC2525 Silicone copolymer surfactant.

APCI

Dabco DC5258 Silicone copolymer surfactant.

APCI

XF-N1586

Newly developed silicone copolymer surfactant, which promotes open cells.

APCI

XF-N1587

Newly developed silicone copolymer surfactant, which promotes open cells.

APCI

DEOA-LF (Dabco)

Diethanolamine Liquid Form (85% DEOA: 15% water)

APCI

Arcol E848

Conventional polyol with an OH# of 31.5

Lyondell Chemical

Arcol E851

43% solids copolymer polyol with an OH# of 18.5 Lyondell Chemical

NC-630

Polyol with an OH# of 31.4

NC-700

41% solids copolymer polyol with an OH# of 21.0 Dow Chemical

Voranol 3512

Polyether polyol with an OH# of 48.3

Dow Chemical

Polyol A

High functionality triol with an OH# of 32.5

Dow Chemical

Polyol B

41% solids copolymer triol with an OH# of 23.8

Dow Chemical

Polyol C

Polyether polyol with an OH# of 28

Dow Chemical

Polyol D

Cell opening polyol

PRC-798

Solvent-based release agent

Dow Chemical

Chem-Trend

APCI: Air Products and Chemicals, Inc. DEOA: Diethanolamine DPG: Dipropylene glycol TEDA: Triethylene diamine

7

Advances in Urethane Science and Technology

1.2.2 Handmix Evaluations 1.2.2.1 Flexible Moulded Foam Handmix Procedure Handmix experiments were carried out using the following procedure. Formulations were blended together for approximately 10 minutes using a mechanical mixer equipped with a 7.6 cm diameter, high shear, mixing blade, rotating at 5000 rpms. Premixed formulations were maintained at 23 ± 1 °C using a low temperature incubator. Mondur TD-80 (Bayer; a blend of 2,4-TDI and 2,6-TDI isomers in the ratio of 4:1) or modified MDI was added to the premix at the correct stoichiometric amount for the reported index for each foam. The mixture was blended together with a Premier Mill Corporation Series 2000, Model 89, dispersator for approximately five seconds. The foaming mixture was transferred to an Imperial Bondware #GDR170 food container or ‘chicken’ bucket and allowed to free rise in order to obtain the processing data.

1.2.2.2 Flexible Slabstock Foam Handmix Procedure Handmix experiments were carried out using the following procedure. A premix consisting of polyol, surfactant and water was prepared by blending the components in a shaker for approximately 20 minutes. The premix was allowed to stand for 2 hours prior to making the foam to allow for degassing of the mixture. A measured amount of premix was poured into a 1.9 litre paper cup; the required stoichiometric amounts of amine and tin catalysts were added to the contents of the cup and mixed for 20 seconds using a Premier Mill Corporation dispersator equipped with a 5.5 cm diameter, high shear, mixing blade, rotating at 6,000 rpm. The corresponding amount of Mondur TD-80 to provide for a 110 index (isocyanate index, which is the amount of isocyanate used relative to the theoretical equivalent amount [5]) was measured into a 400 cm3 beaker. Methylene chloride in the correct proportion was added to the beaker containing the Mondur TD-80; the beaker was carefully swirled for 4 or 5 seconds and the contents poured into the paper cup. The mixture was blended together for 6-7 seconds and the foaming mixture poured into a paper bucket for up to 12 seconds and allowed to free rise with the processing data being recorded. Reactivity profiles were determined from hand-mix foams prepared in 5.68 litre paper buckets. Foams for physical properties were prepared in 35.6 x 35.6 x 25.4 cm cardboard boxes. Identical procedures were followed for both reactivity and physical property experiments.

8

Dimensional Stabilising Additives for Flexible Polyurethane Foams

1.2.3 Machine Evaluation 1.2.3.1 TDI Flexible Moulded Foam Procedure Machine runs for the TDI flexible moulded foam were carried out on a Hi Tech SureShot MHR-50 (Hi-Tech Industries, Inc.), cylinder displacement series, high pressure machine. Fresh premixes, consisting of the appropriate polyols, water, crosslinker, surfactants and catalysts for each formulation were charged to the machine. Mondur TD-80 was used throughout the entire study. All chemical temperatures were held at 23 ± 2 °C via the machine’s internal temperature control units. The foam was poured into an isothermally controlled, heated aluminium mould maintained at 71 ± 2 °C. The mould was a typical physical property tool designed with internal dimensions of 40.6 cm x 40.6 cm x 10.2 cm. The mould has five vents, each approximately 1.5 mm in diameter, centred 10.0 cm from each edge and the geometric centre of the lid. The mould was sprayed with a solvent-based release agent, Chem-Trend PRC-798, prior to every pour and allowed to dry for one minute before pouring. The foam premix was puddle poured into the centre of the mould with a wet chemical charge weight capable of completely filling the mould and obtaining the desired core density. Minimum fill requirements were established for each formulation evaluated. The foam article was demoulded at 240 seconds after the initial pour. After demoulding, the foam was placed through a mechanical crusher, tested for FTC measurements, or left uncrushed and set aside for 24 hour shrinkage measurements described in Section 1.2.3.2c. All foams to be tested in each catalyst set were mechanically crushed 1 minute after demoulding using a Black Brothers Roller crusher set to a gap of 2.54 cm. Crushing was carried out three times on each part, rotating the foam 90 degrees after each pass through the roller. All parts produced for physical testing were allowed to condition for at least seven days in a constant temperature and humidity room (23 ± 2 °C, 50 ± 2% relative humidity). Three to four specimens were produced for any given set of conditions. Four test specimens were die-cut from each foam pad and evaluated for each physical property listed in subsequent data tables. All results were included in calculating averages and standard deviation. Each test was carried out as specified in ASTM D3574 [5]. For each formulation evaluated, duplicate free rise ‘chicken’ buckets were poured at the same shot size to determine overall reactivities and foam shrinkage. Data recorded were cream time (the time between the discharge of the foam ingredients from the mixing head and the beginning of the foam rise [5]), top-of-cup (TOC; the time between the discharge of the foam ingredients from the mixing head and when the centre of the foam reaches the same height as the top of the chicken bucket), string gel (the time between

9

Advances in Urethane Science and Technology pouring of the mixed liquids and the time that strings of viscous material can be pulled away from the surface of the foam when it is touched with a tool [5]), full rise time and final height. The free rise buckets were again tested for final heights after 24 hours. Measurements of height were made using a Mitutoyo height gauge. In addition to all the standard tests, several more unique tests were performed where indicated, and are described in Section 1.2.3.2.

1.2.3.2 Tests 1.2.3.2a Maze Flow Mould Test Description A common type of isothermally heated mould was used to determine the flowability of formulations with each of the catalyst candidates. This maze mould is shown in Figure 1.1. Machine foam was poured into the mould at the top left corner of the open cavity as indicated by ‘pour spot’ on the figure. The lid was then closed and clamped tightly. Foam was allowed to free flow consecutively through each of the five gates for the standard 4

Figure 1.1 Diagram of Maze Flow Mould (Top View)

10

Dimensional Stabilising Additives for Flexible Polyurethane Foams minutes prior to demould. Minimum fill was first determined by completely filling the cavity with little or no extrusion through the vent at the end of the fifth gate. Mathematical reduction of the shot size was performed to obtain the first of three systematically scaled down foam fill weights. This first foam should have a fifth leg (the foam in gate 5 of the maze flow mould, see Figure 1.1) which barely touches the front cavity wall. The second reduction in foam fill weight produced a foam that flowed approximately halfway through the fifth gate. The third reduction in foam fill weight was equivalent to the step change from foam 1 to foam 2. Shot times were held constant for each of the three foam fill weights as compared to the control determined standard shot time in any given solids level formulation. These three foams were weighed for total foam pad and fifth leg weight, and measured for fifth leg length to obtain a range of flow values for each of the experimental catalysts compared to the control additives.

1.2.3.2b Dimensional Stability Test Foam dimensional stability is essentially the result of a balance between external and internal forces. The external forces are defined as the ambient pressure along with any additional applied loads. The internal forces are the strength of the polymer matrix and the internal cell pressure [6]. Basically, if the sum of the internal forces is greater than the external forces, the foam will expand. Consequently, if the sum of the external forces is greater than the internal forces the foam will shrink. Any expansion or shrinkage will impact on the internal and/or external forces until an equilibrium is obtained. It is the internal forces, i.e., cell pressure and strength of the polymer matrix as defined by ‘green strength’ or cure, which will have an impact on the dimensional stability performance of the moulded polyurethane. Dimensional stability can be measured on a freshly demoulded part by determining the amount of force required to open cells, as measured by FTC. FTC measurements were made thirty seconds after demoulding. The foam pad was removed from the mould, weighed and placed in the FTC apparatus. The force detection device is equipped with a 2.2 kg capacity pressure transducer mounted between the 323 cm2 circular plate cross head and the drive shaft. The actual force is shown on a digital display. This device mimics the ASTM D3574, Indentation Force Deflection Test [6] and provides a numerical value of the freshly demoulded foam’s initial hardness or softness. The foam pad was compressed to 50 percent of its original thickness at a cross head velocity of 275 mm per minute with the force necessary to achieve the highest compression cycle recorded in whole Newtons. Several compression cycles were completed. A cycle takes approximately 30 seconds to complete. Values are reported as the FTC value for the foam based on the assumption that the lower the FTC values the better the dimensional stability of the foam. This test requires the foam to be fully cured at demould. A dimensionally stable

11

Advances in Urethane Science and Technology foam will exhibit little or no tendency to shrink after demoulding. Poor dimensional stability can result in numerous defects of the polyurethane article, such as lack of fit of a polyurethane piece to the substrate. These defects will ultimately cause loss of revenue to the polyurethane manufacturers because of increased repair and/or scrap rates. Additionally the degree of cell openness of polyurethane foam can be measured directly by the air flow physical properties of the polyurethane part. Higher air flow values measured for a particular foam would indicate that the foam has less of a tendency to shrink and therefore be more dimensionally stable as compared to a foam with lower air flows. Additionally, higher air flows may also indicate that the foam was much easier to crush-out thereby breaking many of the cell windows. Dimensionally stable foam should reduce scrap and rework by allowing the foam to conform to near its original moulded shape or at least return to its original shape after being crushed.

1.2.3.2c Shrinkage Test An additional dimensional stability evaluation was done with a shrinkage template apparatus (Figure 1.2).

Figure 1.2 Diagram of Shrinkage Template Apparatus

12

Dimensional Stabilising Additives for Flexible Polyurethane Foams It was designed to measure the average foam pad shrinkage of an uncrushed foam. This apparatus consists of two 432 mm long x 432 mm wide x 6.35 mm thick, plexiglass plates, mechanically pinned in each corner with threaded bolts to maintain the plates at a constant 102 mm spacing. The single uncrushed foam from each of the catalyst sets was aged 24 hours prior to being placed between the two plates. Both top and bottom plates each contain 5 mm diameter holes evenly spaced, diagonally from corner to corner, 25 mm apart in an X-shaped pattern. Nineteen holes are contained in each leg of the X, for a total of 37 holes per plate. Measurements were made with a digital caliper by inserting the end down through each hole to just touch the foam surface with the indicated value being recorded. All measurements were normalised to discount the plexiglass plate thickness and subsequently averaged to a single mould cavity and lid value.

1.2.3.2d Time Pressure Release Test Time Pressure Release (TPR) is the opening of the mould during the curing cycle to release the internal pressure and then re-closing for the duration of the cure time [8]. The sudden release of the internally generated pressure bursts the cell windows, thereby improving the crushability of the foam. The tool is opened only a few millimetres and for a specific time. TPR can be applied at any time during the curing cycle, however, care must be taken not to perform the operation too early or too late since surface quality issues may occur. A ‘simulated’ TPR process was carried out during this study, whereby the tool lid was opened approximately 1.5 mm for a three second duration. TPR was applied at various time intervals throughout the evaluation. Two mechanical clasps affixed to the top and bottom halves of the tool precisely controlled the gap opening. These clasps were manually opened and closed at the desired TPR time interval.

1.2.3.2e Pail Test Another test was devised to evaluate foam bulk stability in a free-rise mixed foam. Foams prepared from TDI formulations were poured directly from the machine head into a large open pail (the pail is a common high density polyethylene plastic with approximate dimensions of 365 mm high and 290 mm in diameter) at a targeted mass. Several pours were carried out to ensure equivalent catalyst activity amongst each formulation. Foams were allowed to stand for 24 hours prior to removal from the pail. Each foam was weighed to obtain total individual mass. Subsequently, a 25 mm slice was cut directly through the geometric vertical centre of the foam. Foam slices were examined for cell structure.

13

Advances in Urethane Science and Technology 1.2.3.2f Dynamic Fatigue Test Dynamic Fatigue Constant Pounding testing was carried out using standard testing procedures outlined in ASTM D3574-95 [6]. A 60 minute recovery time and 80,000 cycles were used for each test sample.

1.2.3.2g Fogging Test Gaseous foam emissions were compared in the standard fogging test procedure outlined in SAE J1756 [9]. A standard 7.6 cm thick piece of foam was preconditioned for 48 hours, then tested in the Hart fogging apparatus at 100 °C for 3 hours. Glossmeter readings of the foam were taken after 60 minutes at room temperature.

1.2.3.2h Headspace Analysis Test Foam emissions were also evaluated by cutting equivalent portions, approximately 1 gram, of foam from the geometric centre of the moulded foam articles 60 minutes after demoulding. Each sample was inserted into a 20 cm3 Kimble glass crimp top vial with a Teflon seal. Several vials were sealed without foam to be used as blanks to ensure all emissions had eluted through the gas chromatograph prior to the injection of a second gas sample. All vials were loaded into a Tekmar 7000 Headspace Autosampler tray for sequential heating for 1 hour at 54 °C. After heating and temperature equilibration, the headspace of the vial was sampled and directly injected within a closed loop system onto a Hewlett Packard 5890 Series II Plus gas chromatograph containing a HP-5 (5%-diphenyl/95%dimethylsiloxane copolymer) stationary phase column (30 m, 0.25 mm internal diameter, 1.00 mm film thickness). A standard oven heating profile was used to separate the gas components for detection with a Hewlett Packard 5972 Series Mass Selective Detector. Elution peaks were individually identified by comparison to standard libraries.

1.2.3.3 MDI Flexible Moulded Foam Procedure Machine runs for the MDI flexible moulded foam were conducted on a Krauss-Maffei, cylinder displacement series, high pressure machine. Fresh premixes, consisting of the appropriate polyols, water, crosslinker, surfactants and catalysts for each formulation were charged to the machine. Modified MDI was used throughout the entire study. All chemical temperatures were held at 25 °C ± 2 °C via the machine’s internal temperature control units. Foam pours were made into an isothermally controlled heated aluminium mould maintained at 60 °C ± 2 °C. The mould was a typical physical property tool designed with

14

Dimensional Stabilising Additives for Flexible Polyurethane Foams internal dimensions of 40.6 cm x 40.6 cm x 10.2 cm. The mould has two vents each approximately 1.0 mm in diameter centred 10.0 cm from each edge and the geometric centre of the lid. The mould was sprayed with a solvent-based release agent prior to every pour and allowed to dry for one minute before pouring. The foam premix was puddle poured 15 cm away from the geometric centre of the mould with a wet chemical charge weight capable of completely filling the mould with the appropriate core density. Minimum fill requirements were established for each formulation evaluated. The foam article was demoulded at 300 seconds after the initial pour. Upon demoulding, the foam was placed through a mechanical crusher or tested for FTC measurements. The foams were mechanically crushed 1 minute after demoulding using a roller crusher set to a gap of 3.0 cm. Crushing was carried out three times on each part. All parts produced for physical testing were allowed to condition for at least seven days in a constant temperature and humidity room (23 °C ± 2 °C, 50% ± 2% relative humidity). Three to four parts were produced for any given set of conditions. Four test specimens were die-cut from each pad and evaluated for each physical property listed. All results were included in calculating the averages and standard deviation. Each test was conducted as specified in ASTM D3574 [5]. For each formulation evaluated, free rise cup foams (see Section 1.2.2.1) were poured to determine reactivities and foam shrinkage. Data recorded were gel time, full rise time and final height. The free rise cup foams were tested for final heights and free rise density after 24 hours. Height measurements were carried out using a Mitutoyo height gauge. All experimental formulations reported in this work were matched by rise profile to each control formulation. FTC measurements were conducted 90 seconds after demoulding. The foam pad was removed from the mould, weighed and placed in the FTC apparatus (Instron 4502). The force detection device is equipped with a 5.0 kN capacity pressure transducer. The actual force is shown on a digital display. This device mimics the ASTM D3574, Indentation Force Deflection Test and provides a numerical value of freshly demoulded foams initial hardness or softness. The pad was compressed to 70 percent of its original thickness at a cross head velocity of 380 mm per minute with the force necessary to achieve the highest compression cycle recorded in Newtons. Values are reported as the FTC value for the foam based on the assumption that the lower the FTC values the better the dimensional stability of the foam.

1.3 TDI - Flexible Moulded Additives The automotive industry has placed increased pressures on OEM suppliers to improve their productivity, quality and cost of the polyurethane articles which they produce. Styling changes, complex designs and OEM productivity demands for automotive seats have necessitated the need to produce more open foam in all-water blown TDI- and

15

Advances in Urethane Science and Technology MDI-based systems. Furthermore, changes in polyol technologies designed to improve seat comfort factors can have a negative impact on foam openness and dimensional stability. Specifically, the comfort of high resilience foam may be related to the degree of dimensional stability or foam openness. Flexible moulded polyurethane foam requires mechanical crushing to open foam cells, which in turn prevents shrinkage and improves overall dimensional stability. Current mechanical methods for cell opening consist mainly of roller crushing, vacuum rupture and TPR. However, mechanical methods do not always result in complete or consistent cell opening and require a flexible moulded foam producer to invest in additional machinery. Additionally, if the polyurethane article is not crushed properly, dimensional stability suffers which can cause an increase in repair and scrap rates resulting in a negative impact on the cost of production. A chemical method for cell opening would be preferred.

1.3.1 Dimensional Stability Additives for TDI When producing flexible high resilience foam, it is important to provide a wider TPR window to expand processing latitude and at the same time maintain or improve physical properties. This should result in reduced scrap and/or repair rates, providing improved economics for the polyurethane producer. The most commonly used catalyst and surfactant package for all water blown TDIbased moulded foam production is a blend of Dabco 33-LV and Dabco BL-11, coupled with Dabco DC5043 silicone surfactant. Additionally, the acid-blocked counterparts of these two catalysts, Dabco 8154 and Dabco BL-17, can also be used for the production of high resilience moulded foam. A combination of Dabco 33LV and Dabco BL-17 is used to facilitate a short delay in the reactivity of the polyurethane foaming process. A combination of silicone surfactants, Dabco DC5043 and Dabco DC5169, are utilised to provide good foam stabilisation, improve cell regulation and cell wall drainage. These combinations of catalysts and surfactants served as the control additives to which the experimental additives were compared and contrasted. A newly developed cell opening catalyst, Dabco BL-53, was evaluated to determine its impact on dimensional stability and general processability. Dabco BL-53 affords all the benefits of Dabco BL-11 or Dabco BL-17, with the added advantage of cell opening and slightly delayed initiation times. Dabco BL-53 is not a chemical equivalent for Dabco BL-11 or Dabco BL-17; however, it will provide similar performance. For rapid demoulding systems, it is recommend that Dabco BL-53 be used at 0.12 to 0.22 pphp, with the optimum level at 0.16 to 0.19 pphp, in combination with a Dabco 33-LV level at 0.30 to 0.32 pphp.

16

Dimensional Stabilising Additives for Flexible Polyurethane Foams As polyurethane seating design changed from the relatively simple configurations of the early 1990s to the more complex designs of today, the need to improve cell opening or dimensional stability has intensified. Accordingly, silicone surfactants providing improved foam stabilisation, cell regulation and cell wall drainage were needed to enable polyurethane manufacturers to achieve their production goals. Two experimental silicone surfactants, X-N1586 and X-N1587, were developed to provide open foam and promote dimensional stability. A TDI cushion formulation, with a density of 45 kg/m3 and a TDI back formulation, with a density of 35 kg/m3 were used in the TDI automotive study. All of these formulations were modified accordingly with the appropriate crosslinker, water, and additive levels for the chosen density range. These formulations are shown in Tables 1.2 and 1.3.

Table 1.2 ~45 kg/m3 cushion formulation Formulation

Control

BL53

Exp. SSF

BL53/Exp. SSF

Identification

I

II

III

IV

Components

pphp

pphp

pphp

pphp

Polyol A

68.00

68.00

68.00

68.00

Polyol B

32.00

32.00

32.00

32.00

Water

2.77

2.77

2.77

2.77

Dabco 33LV

0.30

0.30

0.30

0.30

Dabco BL-17

0.18

-

0.18

-

Dabco BL-53

-

0.20

-

0.20

Dabco DC-5043

0.75

0.75

-

-

Dabco DC-5169

0.25

0.25

-

-

X-N1586

-

-

0.3

0.3

X-N1587

-

-

0.7

0.7

Dabco DEOA-LF

1.53

1.53

1.53

1.53

TDI (100 index)

37.3

37.3

37.3

37.3

Exp. SSF: experimental silicone surfactant pphp: parts per hundred polyol

17

Advances in Urethane Science and Technology

Table 1.3 ~35 kg/m3 back formulation Formulation

Control

BL53

Exp. SSF

BL53/Exp. SSF

Identification

V

VI

VI I

VIII

Components

pphp

pphp

pphp

pphp

Polyol A

80.00

80.00

80.00

80.00

Polyol B

20.00

20.00

20.00

20.00

Water

3.27

3.27

3.27

3.27

Dabco 33LV

0.30

0.30

0.30

0.30

Dabco BL-17

018

-

0.18

-

Dabco BL-53

-

0.20

-

0.20

Dabco DC-5043

0.75

0.75

-

0.75

Dabco DC-5169

0.25

0.25

-

-

X-N1586

-

-

0.3

0.3

X-N1587

-

-

0.7

0.7

Dabco DEOA-LF

1.53

1.53

1.53

1.53

TDI (100 index)

42.2

42.2

42.2

42.2

1.3.1.1 Reactivity 1.3.1.1a Handmix Mass-Loss/Rate-of-Rise Formulations used in the mass-loss/rate-of-rise are summarised in Tables 1.2 and 1.3. Surfactant and catalyst additives were changed according to the formulation being studied. Foams were run at the optimum index as they were during the machine study. All experiments were duplicated. Each mixed formulation was poured into a ‘chicken’ bucket equipped with a thermocouple positioned at the centre of the bucket resting on a Mettler PM 30,000 balance. The centre height of the rising foam was recorded every second using a DAPS QA Model #2500 rate-of-rise apparatus. Knowing the foam mass, the rate-of-rise and using the ideal gas law, it is possible to calculate the carbon dioxide generated or trapped over time [10, 11]. Figures 1.3 and 1.4 show height versus time achieved for the cushion and back formulations using the control formulations I and V, as well as the experimental formulations II, III, IV, VI, VII and VIII. The foam height versus time graphs clearly indicate higher rates for control formulations I and V in both cushion and back formulations. Cushion and back

18

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.3 Foam Height versus Time - Cushion Formulation

Figure 1.4 Foam Height versus Time - Back Formulation

formulations, which contained the new additives in formulations II, III, IV, VI, VII and VIII did not achieve the same foam height when compared to the control formulations. Several things could cause the differences observed in the foam heights. First, reactivity rates for control formulations might be faster than the experimental catalyst or surfactants. Second, overall foam stability could be compromised for the experimental catalyst and surfactants. Lastly, carbon dioxide might be diffusing from the reacting polyurethane foam

19

Advances in Urethane Science and Technology at an accelerated rate. All of these possibilities were explored. Data generated and observations will be reported which demonstrate how these new additives promote dimensional stability through increased carbon dioxide diffusion. Temperature profiles for control formulations I and V indicate reaction temperatures which fall in the middle of the temperature profiles compared to the new additives denoted in Figures 1.5 and 1.6. These temperature profiles clearly demonstrate that carbon dioxide

Figure 1.5 Temperature versus Time - Cushion Formulation

Figure 1.6 Temperature versus Time - Back Formulation

20

Dimensional Stabilising Additives for Flexible Polyurethane Foams conversion is occurring at the same rate in the control as in the experimental formulations. The fact that carbon dioxide conversion is occurring at the same rates would not account for the lower foam heights observed in Figures 1.3 and 1.4. In the control formulations, I and V, the amount of carbon dioxide diffused was less than the amount of carbon dioxide diffused using the new additives. Figures 1.7 and 1.8 were generated using the ideal gas law from data generated with the mass-loss/rate-of-rise apparatus. Diffusion

Figure 1.7 Carbon Dioxide Trapped versus Time - Cushion Formulation

Figure 1.8 Carbon Dioxide Trapped versus Time: Back Formulation

21

Advances in Urethane Science and Technology of carbon dioxide from cells as they form in the free rise reaction apparently keeps the reacting foam from reaching the same foam height when using these experimental additives. Reactivity profiles for these three new additives are essentially the same. Data discussed later in Section 9.3.1.1b, Tables 1.4 and 1.5, further supports the fact that there is no

Table 1.4 Machine mix free rise reactivity comparison using cushion formulation Control

BL53

Exp. SSF

BL53/Exp. SSF

I

II

III

IV

Cream (Seconds)

4. 7

4.6

4.9

4.9

Top of Cup (Seconds)

24.1

23.6

23.9

23.4

String Gel (Seconds)

55.4

53.7

54.3

54.2

Full Rise (Seconds)

100.4

103.8

105.3

105.6

Rise: Gel Ratio

1.81

2.01

1.94

1.94

Final Height (mm)

278.4

279.6

280.8

272.6

24 Hour Height (mm)

263.4

265.5

265.5

257.6

5.4

5.0

5.4

5.5

Formulation

% Shrinkage

Table 1.5 Machine mix free rise reactivity comparison using back formulation Control

BL53

Exp. SSF

BL53/Exp. SSF

V

VI

VII

VIII

Cream (Seconds)

4.8

4.5

4.8

4.8

Top of Cup (Seconds)

23.5

22.9

25.4

26.5

String Gel (Seconds)

53.2

52.1

55.6

55.9

Full Rise (Seconds)

109.5

108.7

111.9

111.2

Rise/Gel Ratio

2.06

2.08

2.02

1.99

Final Height (mm)

273.6

275.7

268.3

270.1

24 Hour Height (mm)

245.6

246.9

236.1

235.9

% Shrinkage

10.2

10.4

12.0

12.7

Formulation

22

Dimensional Stabilising Additives for Flexible Polyurethane Foams significant change in reactivity when comparing the control formulations to the experimental formulations.

1.3.1.1b Machine Free Rise Reactivity The machine mix free rise reactivity comparison of all formulations are shown in Tables 1.4 and 1.5. This experimental data illustrates that the overall free rise foam reactivity for both the cushion and back formulations remains relatively the same for the beginning of the reaction. The full rise reactivities in cushion formulations II, III and IV and back formulations VII and VIII start to deviate slightly from the control reference formulations I and V. Percent shrinkage remains fairly consistent within the cushion formulations, I, II, III and IV. Increased foam shrinkage was observed with back formulations VII and VIII. This could be attributed to better cell wall drainage efficiency, providing more open foam and/or an overall increased carbon dioxide diffusion through the polymer network.

1.3.1.2 Foam Physical Properties 1.3.1.2a TPR Effect on Machine Run Moulded Foam FTC When producing polyurethane, manufacturers use some type of mechanical crushing to open cells and insure the polyurethane article does not lose dimensional stability. Several techniques can be used to provide the needed mechanical cell opening. Manufacturers will use TPR, which has been described in Section 1.2.3.2d, along with mechanical roller crushing. Some producers will rely exclusively on the roller crushing and vacuum crushing techniques to provide the mechanical cell opening required. In both cases reducing FTC values and improving foam openness is important for producing polyurethane articles that are dimensionally stable. If TPR is carried out too soon during the polyurethane moulding cycle, the article will collapse (blowout) as indicated in Figure 1.9. This is indicative of the foam being insufficiently cured or lacking enough green strength when TPR was applied. If TPR is conducted too late in the manufacturing process scalloping (concave surface areas on the foam article) and tight foam (insufficient number of open cells within the foam article that causes the hot gas to be trapped and upon cooling forces the entire foam part to shrink) may also occur. When scalloping occurs the foam article must be repaired or scrapped. When tight foam occurs dimensional stability will suffer and there will be a negative impact on physical properties as denoted in Figure 1.10. The foam pad in Figure 1.10 was produced at a 140 second TPR without crushing using formulation V.

23

Advances in Urethane Science and Technology

Figure 1.9 Example of a Collapse/Blowout Moulded Foam (Reproduced with permission from APCI)

Figure 1.10 Example of Tight Foam with Shrinkage (Reproduced with permission from APCI)

24

Dimensional Stabilising Additives for Flexible Polyurethane Foams To understand the benefits of these new additive technologies that provide reduced FTC values, a TPR range from 70 to 150 seconds was run for each of the formulations in Tables 1.2 and 1.3. At a 70 second TPR all formulations suffered blowout since the foam was not sufficiently cured and thus lacked green strength. At an 80 second TPR, no formulations evaluated experienced blowouts or collapse; however, slight distortions and imperfections were evident on the foam surfaces to varying degrees of severity. To complete the TPR window, TPR cycle times were continually ramped up in this fashion to determine the upper limit at which TPR could be applied for each formulation. The upper limit is reached for a given formulation when the foam displays the obvious signs of scalloping and/or ‘dishing’ (concave surface areas of the foam). When this occurs the foam is usually very tight and cannot be used as a functional part. Additionally, parts were produced without utilising TPR during the production cycle in order to compare the difference in foam crushability when TPR is used.

1.3.1.2b Cushion Formulation Machine Evaluation Utilising TPR Cushion control formulation I, which is listed in Table 1.2, was evaluated at a 90-100 second TPR. Initial FTC values of 156 N/323 cm2 for a 90 second TPR and 165 N/323 cm2 for an 100 second TPR were observed. These values were acceptable and produced foam parts of good quality. The new additives in formulations II, III and IV produced maximum initial FTC values of 160 N/323 cm2 at TPR of 90 to 100 seconds. Foam produced at an 80 second TPR for the control formulation I and formulations II, III and IV containing the new additives resulted in minor problems with foam quality, i.e., scalloping. At a 70 second TPR, control and experimental formulations failed because of severe blowout. Figures 1.11 and 1.12 show no significant difference of FTC for all formulations evaluated. When TPR values were increased to 120 seconds for control formulation I, initial values increased to 623 N/323 cm2 and scalloping or foam quality suffered (Figure 1.13). However, increasing the TPR time for cushion formulations II, III and IV to 120 seconds, produced maximum initial FTC values of 205 N/323 cm2 (Figure 1.13). Foam surface quality was very good. Increasing the TPR to 140 seconds increased initial FTC values to a maximum of 543 N/323 cm2 and 534 N/323 cm2 for formulations II and III, respectively (Figure 1.14). Foam quality was still very good. Formulation IV, which utilises a combination of both surfactant and catalyst technologies, achieved a lower initial FTC value at a 140 second TPR of 191 N/323 cm2. This was only slightly higher than the FTC values of the control formulation at a 90 second TPR. Figures 1.11, 1.12, 1.13, and 1.14 show the results of FTC through the entire TPR range applied in this study. These figures clearly demonstrate that use of the new additives can reduce FTC values and maintain dimensional stability over the applied TPR range.

25

Advances in Urethane Science and Technology

Figure 1.11 FTC at 90 TPR - Cushion Formulation

Figure 1.12 FTC at 100 TPR - Cushion Formulation

26

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.13 FTC at 120 TPR - Cushion Formulation

Figure 1.14 FTC at 140 TPR - Cushion Formulation

27

Advances in Urethane Science and Technology Figures 1.15, 1.16, and 1.17 further demonstrate the reduction of FTC values that can be achieved when the new catalyst and surfactant technologies are utilised. These figures clearly show a large difference achieved from plotting initial FTC to the final FTC cycle versus increasing TPR cycle times. Moreover, the lower delta FTC values obtained indicate an improved crush-out capability. The optimum TPR range for control formulation I is 90-100 seconds, while the optimum range for formulations II and III is increased to 120

Figure 1.15 Initial FTC versus TPR Time - Cushion Formulation

Figure 1.16 Delta Difference of 1st to 3rd FTC versus TPR Time - Cushion Formulation

28

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.17 Delta Difference of Initial to Final FTC versus TPR Time - Cushion Formulation

seconds. Formulation IV provided reduced FTC values up to a 140 second TPR. Formulation I did not provide acceptable foam surface quality above 100 second TPR. Formulations II, III and IV provided good surface quality and acceptable physical properties throughout the entire TPR range.

1.3.1.2c Back Formulation Machine Evaluation Utilising TPR Back formulation evaluations were carried out in the same manner as the cushion formulation study. Because of lower solids and higher water content necessary to obtain specific densities and physical properties, this system was more sensitive to processing and TPR range than the cushion formulation. Back formulations V, VI, VII and VIII found in Table 1.3, were all run at TPR times of 70-150 seconds and no TPR. Collapsed foam was encountered at TPR times of 70 seconds for all formulations. FTC values for the back control formulation V had an initial FTC value of 138 N/323 cm2 at a 90 TPR and 160 N/323 cm2 at a 100 second TPR time. Formulations VI, VII and VIII produced a maximum initial FTC value at a 90 and 100 second TPR of 138 N/323 cm2. Control formulation V and experimental formulations VI, VII and VIII provided acceptable FTC values at 90 and 100 second TPR times (Figures 18 and 19). No significant difference in FTC was realised with these formulations at these TPR times. The foam quality at this TPR range was also very good. When TPR times were increased to 120 seconds for control formulation V, initial FTC increased to 645 N/323 cm2. Experimental formulations VI, VII and VIII maintained acceptable initial

29

Advances in Urethane Science and Technology

Figure 1.18 FTC at 90 TPR - Back Formulation

Figure 1.19 FTC at 100 TPR - Back Formulation

FTC values of 334 N/323 cm2, 245 N/323 cm2 and 156 N/323 cm2, respectively, (Figure 1.20). At a 120 second TPR, control formulation V exhibited scalloping while the experimental formulations VI, VII and VIII continued to produce parts with good foam quality. Initial FTC values for the control formulation V increased to 1,045 N/323 cm2 at a 140 TPR and severe shrinkage was observed. FTC values for formulations VI and VII reached maximum values of 649 N/323 cm2, respectively, and 627 N/323 cm2 at a

30

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.20 FTC at 120 TPR - Back Formulation.

Figure 1.21 FTC at 140 TPR - Back Formulation.

140 second TPR (Figure 1.21). Foam produced for these two formulations at a 140 second TPR still produced acceptable quality foam. Formulation VIII achieved lower FTC values than any of the back formulations evaluated at a 140 second TPR range. The initial FTC value for formulation VIII was 405 N/323 cm2 with foam surface quality maintained (Figure 1.21). Figures 1.18, 1.19, 1.20, and 1.21 show the results of FTC values through the entire TPR range utilised in this study.

31

Advances in Urethane Science and Technology Figures 1.22, 1.23, and 1.24 further demonstrate the reduction of FTC values that can be achieved in the back formulations when the new catalyst and experimental silicone surfactant technologies are utilised. At a 90 and 100 second TPR, control formulation V produces good quality foam. When the TPR is increased above 100 seconds foam tightness and surface quality issues arise with the control formulation. Formulations VI, VII and VIII yield good quality foam at all TPR cycle times used in this study. As in the cushion formulation, these lower FTC delta values demonstrate how these newly developed additives enhance crush-out capabilities. Furthermore, this effectively illustrates that the TPR window can be extended as compared to the control formulation.

Figure 1.22 Initial FTC versus TPR Time - Back Formulation

Figure 1.23 Delta Difference of 1st to 3rd FTC versus TPR Time - Back Formulation 32

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.24 Delta Difference of Initial to Final FTC versus TPR Time - Back Formulation

1.3.1.2d Cushion and Back Formulation Machine Evaluation Without TPR Several polyurethane manufactures do not utilise TPR in their production process. The formulations evaluated in this study were designed to utilise TPR when producing polyurethane articles. However, indications of reduced force required to crush a pad using roller crushing and/or vacuum rupture can be understood from the data generated when TPR is not applied to the process. The initial FTC values recorded for the control cushion and back formulations I and V were 1,379 N/323 cm2 and 1,299 N/323 cm2, respectively. Using formulations II, III, VI and VII, FTC values were reduced to 1,152 N/ 323 cm2 and 1,090 N/323 cm2 for the cushion formulation and 1036 N/323 cm2 and 1023 N/323 cm2 for the back formulation. Only minimal shrinkage was observed with the surface quality being maintained. Using formulations IV and VIII, FTC values could further be decreased to 1032 N/323 cm2 and 943 N/323 cm2, respectively. Again, foam surface quality was good with only minimal shrinkage being observed. Figures 1.25 and 1.26 show the FTC values obtained when no TPR was applied during this study.

1.3.1.2e Machine Physical Property Data Comparisons of Various TPR Times Physical properties were evaluated for all the formulations found in Tables 1.2 and 1.3 with TPR being applied at various times throughout the moulding cycle as previously discussed. Additionally, physical properties were evaluated with no TPR being applied. The 90 second TPR time was chosen to be the minimum time to perform TPR for all formulations. Physical property pads were produced at this TPR time. Extended TPR

33

Advances in Urethane Science and Technology

Figure 1.25 FTC at No TPR - Cushion Formulation

Figure 1.26 FTC at No TPR - Back Formulation

times were achieved with the experimental formulations. Physical property pads for physical testing were produced at these extended times. Various TPR windows were identified for the formulations outlined in Tables 1.2 and 1.3. The control formulations I and V both had an upper TPR range of 100 seconds. At a 110 second TPR both formulations I and V had scalloping and surface distortion issues (see Section 1.3.1.2a). Thus the effective TPR range for formulations I and V is 90 to 100 34

Dimensional Stabilising Additives for Flexible Polyurethane Foams seconds. This provides a very narrow process range when utilising TPR with these specific formulations. However, when using the Dabco BL-53 catalyst, or X-N1586/X-N1587 experimental silicone surfactants, both cushion and back formulation TPR times can be extended well beyond the 100 second TPR upper limit of the control formulations I and V.

1.3.1.2f Physical Property Comparison at 90 Second TPR Tables 1.6 and 1.7 provide the physical property comparison for all formulations I-VIII at a 90 second TPR time. The data clearly demonstrates that physical properties are maintained, and in several cases improved, compared to the control formulations. For example, airflow can be improved by as much as 20% as compared to both cushion and back control formulations, when using Dabco BL-53 and experimental silicone surfactants

Table 1.6 Physical properties at 90 TPR: cushion formulation Physical Property

Control

BL53

Exp. SSF

I

II

III

BL53/ Exp. SSF IV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

175

2.2

181

1.5

185

3.1

184

2.8

65%

455

3.5

468

2.1

492

2.4

487

4.1

25% Return

154

1.5

159

1.8

160

1. 5

163

1.8

Ball Rebound (%)

61

(1.5)

63

0.6

61

0.9

62

0.6

Airflow (SLM)

38

5.8

47

4.2

50

5.1

57

6.2

3

Density (kg/m )

44

0.8

44

0. 4

43

0.8

44

0.7

Tensile (kPa)

179

8.4

183

3.1

187

5.4

180

3.2

Tear (N/m)

254

12.1

261

8.9

291

13.3

292

9.4

Elongation (%)

176

7.2

182

6.1

182

5. 8

179

5.1

Wet Set (%)

22

(1.2)

21

0.9

20

0.9

18

1.7

50% Compression Set

5

0.3

5

0.2

5

0.3

4

0.1

50% Humid Aged Compression Set

12

0.9

10

0.4

9

0.3

8

0.5

SD: Standard Deviation ILD: indentation load deflection SLM: standard litres per minute

35

Advances in Urethane Science and Technology

Table 1.7 Physical properties at 90 TPR: back formulation Physical Properties

Control

BL53

Exp. SSF

V

VI

VII

BL53/ Exp. SSF VIII

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

125

3.0

134

3.6

144

1.8

155

1.7

65%

35 4

3.9

370

5.8

396

5.9

422

6.0

25% Return

10 8

1.7

116

2.9

122

1.7

128

3.1

Ball Rebound (%)

60

2.1

60

0.9

60

1.2

60

1.2

Airflow (SLM)

37

2.5

48

1.2

49

4.1

50

3.8

3

Density (kg/m )

35

0.6

36

1.2

36

1.2

36

1.2

Tensile (kPa)

137

6.3

145

5.1

155

9.7

163

10.1

Tear (N/m)

245

9.6

254

10.5

268

11.5

275

11.8

Elongation (%)

165

8.2

176

4.7

170

8.9

178

10.5

Wet Set (%)

24

1.1

22

1.1

18

1.2

19

1.2

50% Compression Set

8

0.5

8

0.2

4

0.2

4

0.5

50% Humid Aged Compression Set

14

1.2

10

0.3

10

0.4

9

0.5

SD: Standard Deviation ILD: indentation load deflection SLM: standard litres per minute

X-N1586/X-N1587 in combination (formulations IV and VIII). Furthermore, Japanese wet set values [6] can be improved by 5 to 8% with the Dabco BL-53, and experimental silicone surfactants X-N1586 and X-N1587. Wet set values are improved nearly 20% when the two additives are utilised in combination as evidenced by Formulations IV and VIII. Equally important, humid aged compression set, tensile and tear physical properties display a positive improvement trend as compared to the control formulations.

36

Dimensional Stabilising Additives for Flexible Polyurethane Foams 1.3.1.2g Physical Property Comparison at 130 and 150 Second TPR Tables 1.8 and 1.9 illustrate the physical properties for selected formulations at TPR times of 130 and 150 seconds. Control physical properties were not evaluated at these extended TPR times since the control formulations I and V had visual surface distortions and severe scalloping. The data in Tables 1.8 and 1.9 demonstrate that the TPR window can be extended by using these newly developed additives without negative impact to the physical properties. In fact, the data indicates that several of the physical properties for the experimental formulations exceed the control formulation properties at the 90 second TPR time. For example, airflow measurements are improved when utilising the extended TPR times. Improvements are greater than 10% with the Dabco BL-53 catalyst and experimental silicone surfactant combinations. Additional improvements are also observed with wet set and 50% humid aged compression set values.

Table 1.8 Physical properties at 130 TPR and 150 TPR: cushion formulation Physical Properties

130 TPR BL53 II

150 TPR Exp. SSF III

BL53/Exp. SSF IV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

25%

175

1.8

164

3.5

174

1.1

65%

434

2.5

457

1.0

467

4.5

25% Return

138

1.8

143

1.5

152

1.1

Ball Rebound (%)

61

0.5

62

0.5

61

0.4

Airflow (SLM)

43

3.2

46

4.7

53

7.4

3

Density (kg/m )

43

0.8

42

0.8

43

0.7

Tensile (kPa)

174

4.1

180

5.1

170

3.5

Tear (N/m)

255

7.4

274

10.5

279

12.1

Elongation (%)

180

3.9

180

3.5

173

3.3

Wet Set (%)

21

1.5

19

1.3

19

0.3

50% Compression Set

5

0.5

5

0. 1

4

0 .3

50% Humid Aged Compression Set

10

0.8

10

0.8

8

0.3

37

Advances in Urethane Science and Technology

Table 1.9 Physical properties at 130 TPR and 150 TPR: back formulation Physical Properties

130 TPR BL53

Exp. SSF

VI

VII

150 TPR BL53/ Exp. SSF VIII

BL53/ Exp. SSF VIII

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

111

2.1

145

4.2

156

0.7

155

1.2

65%

344

5.1

396

3.5

435

8.0

430

3.0

25% Return

96

2.2

123

2.2

132

0.6

129

2.1

Ball Rebound (%)

59

1.4

59

1.5

62

0.7

60

0.2

Airflow (SLM)

49

5.5

45

2.3

49

1 .8

52

2.1

3

Density (kg/m )

35

0.8

36

1.0

36

0.8

36

0.2

Tensile (kPa)

136

5.1

148

5.4

152

7.9

147

6.1

Tear (N/m)

254

9.1

261

8.0

248

8.0

239

5.8

Elongation (%)

170

9.3

166

5.3

163

5.6

175

4.9

Wet Set (%)

22

1.1

19

0.8

19

0.5

19

0.8

50% Compression Set

9

0.2

5

0.3

4

0.2

4

0.6

50% Humid Aged Compression Set

10

0.5

10

0.6

9

0.5

9

0.8

1.3.1.2h Physical Property Comparison with No TPR Tables 1.10 and 1.11 illustrate the physical property performance of all formulations I through VIII without utilising TPR. Given that the specific formulations were designed to be used with the TPR process, foam pads produced under these conditions were very tight upon demoulding. Mechanical crushing at 30 seconds after demould was required instead of the one minute time frame; otherwise, irreversible shrinkage may have occurred prior to placing the foam through the roller crusher. The data in Tables 1.10 and 1.11 indicate that utilising Dabco BL-53 catalyst and/or the experimental silicone surfactants X-N1586 and X-N1587, have a positive influence on the physical properties even when no TPR is applied. Moreover, if the properties of the new additives are compared to the control formulations I and V at the 90 second TPR time (Tables 1.5 and 1.6), physical properties are not adversely affected. In some instances the properties are still maintaining their improvement over the control formulations. Control formulations run at no TPR exhibited surface quality distortions.

38

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.10 Physical properties at no TPR: cushion formulation Physical Properties

Control

BL53

Exp. SSF

I

II

III

BL53/ Exp. SSF IV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

145

0.5

150

1.2

157

0.5

157

2.1

65%

395

0.5

407

2.2

428

3.0

428

4.2

25% Return

127

0.5

131

1.1

136

0.5

138

2.1

Ball Rebound (%)

60

0.7

61

1.4

61

1.0

62

0.7

Airflow (SLM)

34

2.6

42

2.5

47

6.2

49

1.3

3

Density (kg/m )

42

0.8

42

0.5

41

0.6

41

0.6

Tensile (kPa)

159

5.3

160

5.4

175

3.5

169

4.5

Tear (N/m)

241

6.7

252

6.7

269

10.6

268

11.1

Elongation (%)

174

4.1

180

4.2

174

4.5

174

4.3

Wet Set (%)

24

0.6

22

0.9

22

0.2

20

0.3

50% Compression Set

6

0.2

5

0.1

5

0.2

5

0.2

50% Humid Aged Compression Set

12

0.2

10

0.1

10

0.7

9

0.1

1.3.1.2i Physical Property Review All of the physical property data generated at the higher TPR cycle times illustrates that the window for TPR can be extended when using Dabco BL-53 catalyst and X-N1586, XN1587 experimental silicone surfactants. Physical properties were not adversely affected when compared to the control samples. By extending the effective TPR window, polyurethane moulders who utilise this process will enjoy a greater latitude for their processing, providing more freedom to troubleshoot difficult tooling and moulding operations. Additionally, polyurethane manufacturers that do not practice the TPR process can also benefit with improvements realised for physical properties and dimensional stability, as illustrated by many of the physical properties enhancements found in Tables 1.9 and 1.10. Finally, the question remains whether the stability, especially shear stability, suffers from the significantly improved breathability and dimensional stability provided with these new additives. Examination of cellular structures of the machine mixed free rise ‘chicken’ bucket

39

Advances in Urethane Science and Technology

Table 1.11 Physical properties at no TPR: back formulation Physical Properties

Control

BL53

Exp. SSF

V

VI

VII

BL53/ Exp. SSF VIII

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

103

0.5

113

2.1

127

0.7

138

5.0

65%

297

5.0

330

7.7

355

0.6

386

8.4

25% Return

89

0.5

98

2.5

107

0.6

115

3.5

Ball Rebound (%)

58

0.7

61

0.1

61

0.1)

61

1.4

Airflow (SLM)

39

1.1

46

3.1

47

5.1

46

3.8

3

Density (kg/m )

35

0.5

35

0.8

35

0.8

35

0.8

Tensile (kPa)

123

4.9

135

8.8

132

5.7

140

9.9

Tear (N/m)

235

6.5

245

6.3

252

11.6

239

5.9

Elongation (%)

161

8.2

175

7.0

165

6.1

170

9.2

Wet Set (%)

23

1.8

22

1.6

19

0.2

21

1.0

50% Compression Set

9

1.0

9

0.4

6

0.4

5

0.3

50% Humid Aged Compression Set

14

1.8

11

0.5

11

0.6

10

0.6

foams demonstrated no evidence of elongated cells or side shear in the outer edges of the cup. Furthermore, when the TPR ramp study was being carried out it was determined that all formulations had similar lower limit TPR times of approximately 90 seconds. Essentially, this early TPR time would mimic a poorly sealed mould so that the shear forces become obvious in and around the gap or space created by opening the mould. The foam mixture, if not sufficiently stabilised or adequately cured, would be accompanied by a pressure relief which leads to significant mechanical stress, causing collapse. The fact the Dabco BL-53 catalyst, X-N1586 and X-N1587 experimental silicone surfactants did not have blowouts or shear collapses at the 90 second TPR illustrates shear stability is not compromised.

1.3.1.3 Scanning Electron Microscopy A scanning electron microscopy study of moulded foam parts was carried out to understand what affect the additives might have on the morphology of the foam cell structure. It is clear from the SEM photomicrographs in Figures 1.27, 1.28 and 1.29 that there is no 40

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.27 Formulation I, Control (Reproduced with permission from APCI)

Figure 1.28 Formulation II, BL-53 (Reproduced with permission from APCI)

Figure 1.29 Formulation III, Exp. SSF (Reproduced with permission from APCI)

41

Advances in Urethane Science and Technology visual effect on the cell structure of foam produced with the new additives as compared to the control. The additives did not appear to have an affect on the cell morphology of a crushed foam. However, it was not possible to determine differences in cell morphology from an uncrushed part, since the parts were mechanically crushed to prevent normal distortion from foam shrinkage. Formulations I, II and III were used for the SEM study.

1.3.2 Low Emission Dimensional Stability Additives Another emerging need in the automobile market segment is for lower and ultimately no emissions of polyurethane foam additives into the environment. Current efforts revolve around automotive passenger compartment air quality improvements by reduction of the emissions typically from the current migratory additives used to produce interior foam components. These emissions can cause a variety of problems in the final application of the foam, ranging from odour to vinyl staining and window fogging. The migratory nature of current additives is accelerated by sunlight and heat. Thermally induced dehydrohalogenation of the foam/covering interface is assisted by the migration of fugitive tertiary amine catalysts causing vinyl staining (colour change). Breakdown of the vinyl clad covering subsequently releases other chemical compounds, such as plasticisers, which then contribute to window fogging. Consumers and automakers are becoming more conscious of automotive interior air quality, interior vinyl degradation and film formation (windshield fogging) [12]. This section of the chapter reports on the development of non-fugitive gelling and blowing polyurethane catalysts to address the aforementioned emission issues. These catalysts chemically bind to the polyurethane foam matrix (contain active hydrogens) rendering them incapable of migrating back out of the foam after the reaction is complete, while still producing a quality product. Equivalent or improved physical properties such as airflow, compression set and crushability, must be maintained as compared to the industry standards used in TDI-, MDI-based systems, and TDI/MDI blend systems. Presented in this section is work using these non-fugitive catalysts in a variety of flexible polyurethane foam applications, including MDI and TDI automotive and flexible slabstock formulations. The results show that these new additives can give dimensional stability and processability to the foam, while providing non-fugitive and non-fogging benefits. Additionally, improvements to the physical properties of the polyurethane article when using the new additives, occurs. Accordingly, these non-fugitive catalysts provide good to excellent foam stabilisation, cell regulation and cell wall drainage and are needed to enable polyurethane manufacturers to maintain their standard production goals. Two experimental non-fugitive catalysts, XF-N1085 and XF-O11006 were developed to provide open foam and promote

42

Dimensional Stabilising Additives for Flexible Polyurethane Foams dimensional stability. During the development of non-fugitive catalysts, it was observed that TDI formulations, in general, tend to be more closed cell with the use of these catalysts compared to MDI systems. Therefore, the technology was advanced to incorporate non-fugitivity and cell opening characteristics within a single catalyst. XF-N1085 is a novel cell opening blowing, non-fugitive catalyst which affords all the benefits of the well known Dabco BL-11, with the added advantage of cell opening and no foam emissions. XF-N1085 is not a drop in replacement for Dabco BL-11. For rapid demoulding of TDI systems, it is recommended that XF-N1085 be used at 0.2 to 0.4 php, with the optimum level at 0.25 to 0.35 pphp, in combination with an XF-O11006 level at 0.6 to 0.9 pphp. XF-O11006 is a novel cell opening gelling non-fugitive catalyst which gives all the benefits of the well known Dabco 33-LV, with the added advantage of cell opening, low volatility and no foam emissions. XF-O11006 is not a drop in replacement for Dabco 33-LV. Several representative industry cushion and back formulations, listed in Tables 1.121.15, were utilised to compare multiple additives. Dabco NE200 was compared against

Table 1.12 ~45 kg/m3 TDI Lyondell cushion formulation Formulation identification

IX

X

XI

XII

Components

pphp

pphp

pphp

pphp

Arcol E848

50.00

50.00

50.00

50.00

Arcol E851

50.00

50.00

50.00

50.00

Water

2.34

2.34

2.34

2.34

Dabco 33LV

0.25

-

-

-

Dabco BL-11

0.10

-

-

-

Dabco NE1060

-

0.08

0.53

-

Dabco NE200

-

0.64

-

-

XF-N1085

-

-

0.26

0.26

XF-O11006

-

-

-

0.68

Dabco DC5043

0.75

0.75

0.75

0.75

Dabco DEOA-LF

1.76

1.76

1.76

1.76

TDI (100 index)

32.74

32.74

32.74

32.74

43

Advances in Urethane Science and Technology

Table 1.13 ~35 kg/m3 TDI Lyondell back formulation Formulation Identification

XIII

XIV

XV

XVI

Components

pphp

pphp

pphp

pphp

Arcol E848

70.00

70.00

70.00

70.00

Arcol E851

30.00

30.00

30.00

30.00

Water

3.24

3.24

3.24

3.24

Dabco 33LV

0.25

-

-

-

Dabco BL-11

0.10

-

-

-

Dabco NE1060

-

0.08

0.53

-

Dabco NE200

-

0.64

-

-

XF-N1085

-

-

0.26

0.26

XF-O11006

-

-

-

0.68

Dabco DC5043

1.00

1.00

1.00

1.00

Dabco DEOA-LF

1.76

1.76

1.76

1.76

TDI (100 index)

41.87

41.87

41.87

41.87

Table 1.14 ~45 kg/m3 TDI Dow cushion formulation Formulation identification

XVII

XVIII

XIX

XX

Components

pphp

pphp

pphp

pphp

NC-630

46.00

46.00

46.00

46.00

NC-700

54.00

54.00

54.00

54.00

Water

2.34

2.34

2.34

2.34

Dabco 33LV

0.25

-

-

-

Dabco BL-11

0.10

-

-

-

Dabco NE1060

-

0.22

0.70

-

Dabco NE200

-

0.64

-

-

XF-N1085

-

-

0.27

0.25

XF-O11006

-

-

-

0.80

Dabco DC5164

1.0

1.0

1.0

1.0

Dabco DC5169

0.5

0.5

0.5

0.5

Dabco DEOA-LF

1.76

1.76

1.76

1.76

TDI (100 index)

32.70

32.70

32.70

32.70

44

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.15 ~35 kg/m3 TDI Dow back formulation Formulation identification

XXI

XXII

XXIII

XXIV

Components

pphp

pphp

pphp

pphp

NC-630

68.00

68.00

68.00

68.00

NC-700

32.00

32.00

32.00

32.00

Water

3.24

3.24

3.24

3.24

Dabco 33LV

0.25

-

-

-

Dabco BL-11

0.10

-

-

-

Dabco NE1060

-

0.27

0.58

-

Dabco NE200

-

0.52

-

-

XF-N1085

-

-

0.32

0.24

XF-O11006

-

-

-

0.77

Dabco DC5164

0. 2

0.2

0.2

0.2

Dabco DC5169

0.6

0.6

0.6

0.6

DEOA-LF

1.76

1.76

1.76

1.76

TDI (100 index)

41.92

41.92

41.92

41.92

Dabco BL-11 for equivalent blowing efficency. Dabco NE1060 was compared against Dabco 33LV for equivalent gelling efficiency. Required blow:gel ratios are listed in each table. Additionally, XF-N1085 and XF-O11006 were specifically developed as dimensional stability/cell opening additives and they are also compared here.

1.3.2.1 Reactivity Tables 1.16 and 1.17 illustrate that the overall free rise foam reactivity for both the cushion and back formulations remains relatively the same for the entire foaming reaction. This was achieved by balancing the experimental catalyst levels, blow:gel ratios, and matching them to the control Dabco BL-11 and Dabco 33LV levels in IX, XIII, XVII and XXI.

45

Advances in Urethane Science and Technology

Table 1.16 Machine mix free rise reactivity comparison using TDI cushion formulations Formulation

IX

X

XI

XII

XVII

XVIII

XIX

XX

Cream (Seconds)

4.9

4.8

4.8

4.5

3.5

3.8

4.5

4.9

Top of Cup (Seconds)

37

36

39

35

32

35

36

37

String Gel (Seconds)

56

55

56

53

53

53

55

56

Full Rise (Seconds)

83

81

84

78

71

75

81

83

Table 1.17 Machine mix free rise reactivity comparison using TDI back formulations Formulation

XIII

XIV

XV

XVI

XXI

XXII

XXIII

XXIV

Cream (Seconds)

3.8

3.8

3.8

3.4

4.2

3.2

3.2

3.2

Top of Cup (Seconds)

30

33

32

31

32

37

36

37

String Gel (Seconds)

58

56

59

57

54

55

57

56

Full Rise (Seconds)

82

85

94

91

75

86

85

88

1.3.2.2 Flexible Moulded Foam Machine Physical Property Data To understand the benefits of these new additive technologies that provide low volatility, no amine emissions, no fogging and, in some cases, cell opening, all TDI moulded formulations were run on the Hi Tech machine. Several foams were produced for each catalyst combination and formulation to obtain physical property pads, FTC pads, shrinkage pads, and flow evaluations.

46

Dimensional Stabilising Additives for Flexible Polyurethane Foams 1.3.2.2a Standard Physical Properties Tables 1.18 to 1.21 provide the physical property comparison for the TDI flexible moulded formulations IX-XXIV. The data clearly demonstrates that physical properties are maintained, and in several cases improved, compared to the control formulations, depending on the formulation and the use of experimental cell-opening catalysts. For example, ILD properties were comparable or improved over the control in both low and high solids formulations, especially in the Dow polyol system. This is attributed to the slightly different polymer matrix resulting from non-fugitive catalysis which can produce higher load properties. In almost all cases, the support factor was comparable to the control. Tensile values in the Dow system were matched to the control using XF-N1085 in combination with Dabco NE1060, formulations XIX and XXIII. Dry and humid compression sets (50% dry, 50% humid aged, and 75% humid aged) were, in many formulations, slightly elevated, with the Dabco NE1060/XF-N1085 catalyst combination (XIX and XXIII) closest to the control. The compression set data illustrates that this

Table 1.18 Physical properties for Lyondell TDI cushion formulations Physical Property

IX

X

XI

XII

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

180

-

170

-

182

-

181

-

65%

534

-

486

-

522

-

517

-

25% Return

157

-

150

-

160

-

158

-

Ball Rebound (%)

64

1.0

65

0.8

64

1.2

65

0.7

Airflow (SLM)

48

5.7

45

8.5

62

11.3

40

2.8

3

Density (kg/m )

42

1.3

44

0.5

43

1.0

44

1.0

Tensile (kPa)

143

10.1

136

66.7

163

7.8

176

6.5

Tear (N/m)

193

12

186

7

217

23

193

7

Elongation (%)

98

7.0

102

6.3

97

6.1

102

10.6

Wet Set (%)

18

1.0

20

1.2

20

0.9

22

1.4

50% Compression Set (%)

5

0.4

8

1.1

6

0.1

7

0.7

50% Humid Aged Compression Set (%)

9

0.2

16

0.8

6

0.5

17

1.2

Hysteresis (%)

19

0.1

18

0.1

19

0.1

20

0.4

Support Factor

3.0

-

2.9

-

2.9

-

2.9

-

47

Advances in Urethane Science and Technology

Table 1.19 Physical properties for Lyondell TDI back formulations Physical Property

XIII

XIV

XV

XVI

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

103

-

97

-

99

-

103

-

65%

315

-

287

-

289

-

310

-

25% Return

87

-

84

-

86

-

89

-

Ball Rebound (%)

62

0.8

64

0.56

64

1.1

62

0.7

Airflow (SLM)

65

5.7

62

8.5

71

11.3

59

2.8

Density (kg/m3)

31

1.0

30

0.9

31

1.3

30

0.9

Tensile (kPa)

104

4.5

119

6.8

112

8.7

121

4.4

Tear (N/m)

243

19

224

21

259

16

259

19

Elongation (%)

93

5.7

97

7.6

100

4.2

100

7.4

Wet Set (%)

27

0.9

25

0.9

26

1.0

26

0.9

50% Compression Set (%)

6

0.7

8

0.4

7

0.53

8

0.4

50% Humid Aged Compression Set (%)

16

0.6

20

1.2

20

1.1

17

1.0

Hysteresis (%)

21

0.1

18

0.2

20

0.2

20

0.1

Support Factor

3.1

-

3.0

-

2.9

-

3.0

-

property is very system dependent when using non-fugitive catalysts. For performance optimisation, one may want to consider reformulation to produce the desired physical properties. The use of XF-N1085 and XF-O11006 will improve or maintain physical properties compared to the control for TDI systems. The Japanese wet set data was determined to be comparable or, in the case of the Lyondell back formulation, slightly improved when compared to the control foam. The use of cell opening non-fugitive catalysts was found to improve wet sets in the Lyondell high solids formulation. In addition, the cell opening catalysts were also found to decrease hysteresis and increase tensile strength in the back formulations using Lyondell polyol compared to the control formulation.

48

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.20 Physical properties for Dow TDI cushion formulations Physical Property

XVII

XVIII

XIX

XX

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

106

-

180

-

180

-

193

-

65%

321

-

51 3

-

518

-

551

-

25% Return

92

-

155

-

156

-

166

-

Ball Rebound (%)

56

0.6

55

1.9

56

0.8

57

0.8

Airflow (SLM)

17

2.8

23

5.7

17

2.8

20

2.8

Density (kg/m3)

45

0.4

44

0.6

44

0.8

44

0.7

Tensile (kPa)

215

14.3

176

6.8

205

4.0

171

13.7

Tear (N/m)

210

11

210

16

235

18

221

14

Elongation (%)

100

7.0

93

6.6

89

6.9

103

5.8

Wet Set (%)

16

0.2

21

1.4

20

0.6

17

0.6

50% Compression Set (%)

6

0.6

9

0. 1

8

0.3

8

0.3

50% Humid Aged Compression Set (%)

13

0.5

17

0.8

20

2.6

18

0.7

Hysteresis (%)

21

0.1

21

0.1

21

0.6

22

0.1

Support Factor

3.1

-

2.9

-

2.9

-

2.9

-

1.3.2.2b Dynamic Fatigue Results Table 1.22 shows the dynamic fatigue test results for Lyondell and Dow cushion foams tested for 80,000 cycles and 60 minute recovery time. Results indicate that equivalent dynamic fatigue values are obtainable with use of the new non-fugitive and cell opening catalysts, based on nominal test error of ± 3.

1.3.2.2c FTC Results In both the Lyondell cushion and back formulations IX through XVI listed in Tables 1.11 and 1.12, addition of XF-N1085 and XF-O11006 result in a reduction in FTC values (Tables 1.22 and 1.23) throughout the initial several repetitions of the crushing cycle illustrated in Figures 1.30 and 1.31. At the end of the ten complete repetitions, all three experimental catalyst packages, resulted in the same final FTC value. Without the

49

Advances in Urethane Science and Technology

Table 1.21 Physical properties for Dow TDI back formulations Physical Property

XXI

XXII

XXIII

XXIV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

108

-

100

-

110

-

115

-

65%

315

-

299

-

313

-

311

-

25% Return

94

-

87

-

95

-

100

-

Ball Rebound (%)

63

0.8

60

1.7

60

1.9

63

0.8

Airflow (SLM)

48

2.8

62

8.5

45

5.7

59

5.7

Density (kg/m3)

33

1.0

33

0.4

33

0.9

32

0.3

Tensile (kPa)

145

5.4

127

7.4

141

5.8

139

4.8

Tear (N/m)

249

9

242

4

247

16

263

23

Elongation (%)

106

11.0

98

8.0

109

12.7

105

4.9

Wet Set (%)

21

0.9

16

1.3

22

1.7

19

1.9

50% Compression Set (%)

5

0.4

7

0.6

8

0.6

7

0.4

50% Humid Aged Compression Set (%)

15

0.9

31

2.1

21

0.2

26

1.8

Hysteresis (%)

19

0.1

18

0.2

20

0.1

19

0.2

Support Factor

2.9

-

3.0

-

2.9

-

2.7

-

Table 1.22 Dynamic fatigue ASTM D3574-95 [5] (80,000 cycles) Original 40% Deflection Force (N)

Final 40% Deflection Force (N)

% Change (%)

IX

282

244

13.4

X

258

207

19.9

XI

273

222

18.6

XII

272

238

12.7

XVII

302

228

24.4

XVIII

261

206

20.9

XIX

264

200

24.3

XX

280

297

29.5

50

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.23 FTC values for TDI cushion in Lyondell system Lyondell High Solids (N) Frequency

FTC Value

FTC Value

FTC Value

FTC Value

IX

X

XI

XII

1

767

923

738

698

2

356

540

345

349

3

242

316

234

249

4

198

236

189

205

5

167

193

162

178

6

149

167

160

171

7

140

151

158

167

8

138

145

156

160

9

138

140

158

160

10

133

127

156

160

Figure 1.30 FTC Graph for TDI Cushion in Lyondell System (Ref. Table 1.22)

51

Advances in Urethane Science and Technology

Figure 1.31 FTC Graph for TDI Back in Lyondell System (Reference Table 1.23)

use of these cell opening catalysts, the Dabco NE1060/Dabco NE200 non-fugitive catalyst package tightens the foam resulting in slightly elevated values during several initial FTC footplate depressions. In the cushion Dow formulations XVII through XX, the XF-N1085 and XF-O11006 show similar trends to the Lyondell formulations with reduced FTC (Figure 1.32 and Table 1.24), thereby producing a more dimensionally stable foam. In contrast, for the low solids back formulations XXI through XXIV, all catalyst combinations exhibited no significant difference to the control catalyst (Figure 1.33 and Table 1.25). Inherent tightness in this particular back formulation did not allow the same degree of catalyst differentiation to occur in these foams. Adjustment to the catalyst loading levels would be required to demonstrate the same reduced FTC performance as previously described in other formulations. Figures 1.34 and 1.35 further demonstrate the reduction of FTC values that can be achieved with these new catalyst technologies. The graphs indicate the results achieved by plotting the delta of the initial and third FTC value for every formulation. Moreover, lower delta values illustrate an improved crush-out capability.

52

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.32 FTC Graph of TDI Cushion in Dow System (Reference Table 1.24)

Table 1.24 FTC values for TDI back in Lyondell system Lyondell Low Solids (N) Frequency

FTC Value

FTC Value

FTC Value

FTC Value

XIII

XIV

XV

XVI

1

830

858

698

683

2

474

558

427

409

3

294

360

251

245

4

196

236

169

165

5

151

182

129

129

6

118

136

107

105

7

100

11 6

91

91

8

85

96

80

80

9

76

85

76

71

10

73

80

69

64

53

Advances in Urethane Science and Technology

Figure 1.33 FTC Graphs for TDI Back in Dow System (Reference Table 1.25)

Table 1.25 FTC values for TDI cushion in Dow system Dow High Solids (N) Frequency

FTC Value

FTC Value

FTC Value

FTC Value

XVII

XVIII

XIX

XX

1

1014

1234

959

843

2

654

885

620

516

3

471

641

454

383

4

369

503

360

309

5

314

414

305

267

6

274

347

280

240

7

254

298

256

225

8

236

247

229

211

9

216

249

207

196

10

207

220

187

191

54

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.26 FTC values for TDI back in Dow system Dow Low Solids (N) Frequency

FTC Value

FTC Value

FTC Value

FTC Value

XXI

XXII

XXIII

XXIV

1

936

881

901

905

2

500

505

516

500

3

309

311

311

298

4

227

231

211

209

5

171

178

158

167

6

138

140

127

138

7

116

111

107

118

8

100

96

93

105

9

87

82

82

98

10

80

69

76

89

Figure 1.34 Delta Difference Graph of First and Third FTC Values in Lyondell System

55

Advances in Urethane Science and Technology

Figure 1.35 Delta Difference Graph of First and Third FTC Values in Dow System

1.3.2.2d Pail Test Results Visual examination of the foam slice’s cellular structure showed no evidence of elongated cells or side shear collapse near the outer edges. The appearance of these elongated cells or side shear collapse is a phenomenon created by the frictional drag of the plastic pail (see Section 1.2.3.2e) on the rising foam. Each slice was measured for maximum height, normalised by weight and displayed in Figure 1.36 for comparison of final foam recession to the control. It is evident from the data that no significant difference exists in bulk stability when using any of the cell opening or non-fugitive catalysts.

1.3.2.2e Maze Mould Test Results Data in the Lyondell system, formulations IX through XVI, generated using the previously described maze flow mould is summarised in Tables 1.27 and 1.28 and

56

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.36 Graph Comparing 24 Hour Pail Height

Table 1.27 Maze flow mould data for Lyondell TDI cushion formulations IX - XII Formulation Identifier

Pad wt. (g)

Flow past 5th gate (g)

% of Total pad wt.

MINIMUM 5th Gate IX

405

30

8.0

X

404

30

8.0

XI

405

35

9.5

XII

409

30

7.9

MEDIAN 5th Gate IX

474

101

27.1

X

472

104

28.3

XI

478

109

29.5

XII

479

105

28.1

MAXIMUM 5th Gate IX

524

150

40.1

X

526

160

43.7

XI

531

161

43.5

XII

528

155

41.6

57

Advances in Urethane Science and Technology

Table 1.28 Maze flow mould data for Lyondell TDI back formulations XIII-XVI Formulation Identifier

Pad wt. (g)

Flow past 5th gate (g)

% of Total pad wt.

MINIMUM 5th Gate XIII

619

92

17.5

XIV

617

102

19.8

XV

615

108

21.3

XVI

616

107

21.0

MEDIAN 5th Gate XIII

667

143

27.3

XIV

666

153

29.8

XV

665

160

31.7

XVI

668

156

30.5

MAXIMUM 5th Gate XIII

718

193

36.8

XIV

715

202

39.4

XV

716

206

40.4

XVI

718

208

40.8

shown graphically in Figure 1.37. The graph utilises a xy scatter plot with linear regression trendlines and R2 values (coefficient of determination) reported for each formulation group. It compares estimated and actual y-values and ranges in value from zero to one. Since these values are very close to one, there is a near perfect correlation in the samples. However, if the coefficient of determination approaches zero, the regression equation is not helpful in predicting a y-value. The data trends suggest that increasing part weight, regardless of formulation type, shows no significant penalty in flow compared to the control formulations IX and XIII. Similar maze flow mould results are also obtained with the Dow cushion and back formulations summarised in Tables 1.28 and 1.29 and shown graphically in Figure 1.38. In these Dow formulations, XVII through XXIV, the data suggests slightly improved linear regression trendlines and R2 values. Again, no significant penalty in flow is observed as compared to the control formulations XVII and XXI.

58

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.37 Maze Flow Mould Graph for Formulations IX-XVI

Table 1.29 Maze flow mould data for Dow TDI cushion formulations XVII-XX Formulation Identifier

Pad wt. (g)

Flow past 5th gate (g)

% of Total pad wt.

MINIMUM 5th Gate XVII

404

40

11.0

XVIII

403

36

9.8

XIX

399

34

9.3

XX

404

25

6.6

MEDIAN 5th Gate XVII

457

93

25.6

XVIII

457

92

25.2

XIX

458

100

27.9

XX

453

69

18.0

MAXIMUM 5th Gate XVII

509

149

41.4

XVIII

508

140

38.0

XIX

508

149

41.5

XX

513

136

36.1

59

Advances in Urethane Science and Technology

Table 1.30 Maze flow mould data for Dow TDI back formulations XXI-XXIV Formulation Identifier

Pad wt. (g)

Flow past 5th gate (g)

% of Total pad wt.

MINIMUM 5th Gate XXI

568

65

12.9

XXII

566

70

14.1

XXIII

565

64

12.8

XXIV

569

83

17.1

MEDIAN 5th Gate XXI

617

122

24.7

XXII

616

120

24.2

XXIII

614

112

22.3

XXIV

618

132

27.2

MAXIMUM 5th Gate XXI

668

168

33.6

XXII

665

172

34.9

XXIII

666

164

32.7

XXIV

670

183

37.6

Figure 1.38 Maze Flow Mould Graph for Formulations XVII-XXIV

60

Dimensional Stabilising Additives for Flexible Polyurethane Foams 1.3.2.2f Shrinkage Test Results Shrinkage results displayed in Figures 1.39 and 1.40 were generated with the shrinkage apparatus previously described in the experimental section. Overall, the cavity shrinkage is much greater as compared to the lid shrinkage, evidenced in both Lyondell and Dow polyol systems. The use of Dabco NE1060/Dabco NE200 produced foams with more shrinkage than the control catalysts. However, utilising non-fugitive cell opening catalysts, XF-N1085 and XF-O11006 resulted in shrinkage comparable to the control and in some cases even less. As expected, less shrinkage was observed in both Dow and Lyondell cushion formulations due to the increased solids loading as compared to their respective back formulations.

Figure 1.39 Uncrushed Foam Shrinkage for Lyondell System

Figure 1.40 Uncrushed Foam Shrinkage for Dow System

61

Advances in Urethane Science and Technology 9.3.2.2g Fogging Test Results Several machine foam samples were tested in accordance with fogging test method SAE J1756-94 [9]. Results are listed in Table 1.31. The results obtained in the Lyondell polyol system showed improved percentage reflectance over the control indicating reduced emissions from the foam. Comparable results were obtained in all the Dow polyol system samples regardless of catalyst use.

Table 1.31 Fogging results System Identifier

% of Reflectivity

IX

87

X

97

XI

98

XII

96

XVII

93

XVIII

92

XIX

93

XX

93

1.3.2.2h Headspace Analysis and Vinyl Staining Test Results Headspace analysis in series with gas chromatography and mass spectrometry has indicated no emissions from the experimental non-fugitive catalysts mentioned in this paper. Emissions from the fugitive control catalysts were measured and identified. These results support the fact that these experimental catalysts will result in no VOC emissions which will ultimately yield no amine fogging and no vinyl staining from the final polyurethane foam article. Additional volatility measurements have indicated that these experimental catalysts volatilise 3 to 10 times less than the current industry standards (Dabco 33LV and Dabco BL-11) in the temperature range of 0 °C to 150 °C. Vinyl staining tests using commercial formulations and various grades and types of vinyl have indicated that the experimental catalysts mentioned in this chapter do not contribute to vinyl staining. Compared to industry standard amine catalysts, use of these experimental non-fugitive catalysts will result in significant reduction (ΔE of less than 2) in staining measurements [12]. DE is the numerical total colour difference, using lightness and chromaticity factors, between a sample and a known colour standard.

62

Dimensional Stabilising Additives for Flexible Polyurethane Foams 1.3.2.2i TDI Flexible Moulded Foam Review The most commonly used catalyst packages for all water blown TDI flexible foam production are blends of Dabco 33LV and Dabco BL-11 at a typical ratio of approximately 3:1. These new non-fugitive catalysts, Dabco NE1060, Dabco NE200, XF-N1085, and XF-O11006 are recommended for TDI formulations at various gel:blow ratios depending upon the selectivity desired. Ratios from 2:1 to 8:1 can be used to tailor the reaction profile to the preferred selectivity for optimisation of physical properties and adaptation to various processing conditions. The physical properties obtained with these new nonfugitive catalysts are similar to the physical properties generated with standard industry catalysts. However, slight formulation adjustments may be necessary to improve dry and wet compression set values in certain instances. Improved crush-out capability can also be achieved when incorporating the new cell-opening non-fugitive catalysts, XF-N1085 and XF-O11006. Complementary methods of emission analysis, headspace and fogging, both suggest reduced emissions from the foam with these new non-fugitive catalysts.

1.4 MDI Flexible Moulded Foam Additives The current trend toward lower foam densities, faster demould times, and the increased use of complex metal and plastic insert frames in the construction of automobile interior components has increased the difficulty of controlling the numerous process and chemical variations a polyurethane manufacturer encounters on a daily basis. To successfully produce moulded high resilience polyurethane foam, the manufacturer must maintain a critical balance between foam over stabilisation, resulting in foam shrinkage, and under stabilisation, resulting in internal defects, such as basal cell formation and shear collapse. This often difficult to achieve balance is referred to as processing latitude. Variations on either side of the processing latitude can result in costly increases for the producer in scrap and repair rates. To help minimise production losses, polyurethane manufacturers are essentially forced to over stabilise their foam formulations on a regular basis. Subsequently, the closed cells which normally result from this over stabilisation are opened via standard mechanical crushing techniques [13]. New surfactant technologies have been developed which promote improved cell wall drainage and contribute to the final level of cell openness of the foam product without causing any of the aforementioned negative attributes typical to a foam process. Due to this unique cell opening action of these new additive technologies, the foam crushing portion of the process can be greatly simplified or potentially eliminated. Figure 1.41 shows a variety of flexible moulded Dabco surfactants which can be utilised in TDI and/or MDI formulations. Using Figure 1.41, proper selection of the surfactant can be made, enabling polyurethane formulators to expand their processing latitude, cell openness, and bulk stability of the foam article in their critical formulations.

63

Advances in Urethane Science and Technology

Figure 1.41 Flexible Moulded Surfactant/Property Relationships

1.4.1 Dimensional Stability Additives for MDI An increasing North American trend in the moulded polyurethane foam industry today is the use of MDI. While the advantages and disadvantages of MDI versus TDI are still being debated, the fact remains that most polyurethane producers are using MDI in at least part of their daily flexible foam production. The surfactant requirements for MDIbased formulations can vary dramatically, dependent upon the type of MDI and polyols being used. Most MDI formulations are inherently quite stable; therefore minimal stabilising contributions are necessary from the surfactant. Conventional TDI high resilience moulded surfactants are far too potent for MDI formulations. For example, utilisation of Dabco DC5043 or Dabco DC5164 in an MDI seating formulation may result in over stabilisation and significantly contribute to the shrinkage of the foam article. To demonstrate the effectiveness of these dimensional stabilising additives, Dabco DC2517 and Dabco DC5244, several moulded high resilience foam experiments were carried out using the formulation illustrated in Table 1.32, to compare with two non-cell opening surfactants Dabco DC2585 and an additional surfactant (control). Much of the work described next utilises proprietary formulations and only a general description of some of the components can be disclosed.

64

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.32 MDI flexible moulded formulation Components

pphp

Polyether polyol, diol/triol blend

100

Water

3.5

Diisopropanol amine

2.0

Dabco 33LV

0.3

Dabco BL-11

0.25

Dabco B-16

0.3

Surfactants

1.0

MDI prepolymer (2.3 functionality, 26% NCO)

100 Index

1.4.1.1 Reactivity - Handmix Rate of Rise Data generated in the rate-of-rise comparison tests are summarised in Table 1.33. Foams were run at an optimum index of 100 and all experiments were duplicated. Each mixed formulation was poured into a ‘chicken’ bucket equipped with a thermocouple positioned at the centre of the bucket resting on a Mettler PM 30 balance. The centred height of the rising foam was recorded in millimetres every second using a DAPS (data acquisition and plotting system) QA Model #2500 rate-of-rise aparatus. Reactivity profiles for these new additives are essentially the same. Data discussed next further supports the fact that there is no significant impact on reactivity or physical properties when comparing the control to the Dabco DC2517 and Dabco DC5258 formulations.

Table 1.33 Handmix rate of rise reactivity comparison Formulation Profile Top of Cup 1 (seconds)

12.4

Top of Cup 2 (seconds)

48.1

String Gel (seconds)

53.4

Full Rise (seconds)

101.8

Full Rise Height (mm)

395.1

65

Advances in Urethane Science and Technology

1.4.1.2 Standard Physical Properties To understand the benefits of these new additive technologies that provide low volatility and cell opening, all MDI moulded formulations were run in handmix foam to demonstrate performance. Several foams were produced for each surfactant formulation to obtain physical property pads, FTC pads, shrinkage pads, and flow evaluations. Table 1.34 illustrates moulding foams at a density of 47-48 kg/m3; similar physical properties are obtained for ILD, resilience, airflow, 50% dry sets and extensive properties. However, slight nominal improvements in 50% humid aged compression sets (HACS) and wet sets are achieved with Dabco DC2517 and Dabco DC5258.

Table 1.34 Physical properties for MDI moulded formulations Physical Property

Control

DC2517

DC2585

DC5258

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

302

3.4

295

4.5

292

5.2

300

1.8

65%

764

9.9

762

8.2

759

7.2

753

1.3

25% Return

223

3.5

213

4.7

212

3.9

219

1.0

Ball Rebound (%)

44.3

0.4

43.7

0.7

43.7

0.7

44.0

0.5

Airflow (SLM)

22.5

2.1

24.3

9.1

23.3

5.0

28.1

10.4

Tensile (kPa)

165

5.1

171

9.8

165

11.7

167

8.6

Tear (N/M)

305

11.5

303

15.7

309

11.2

330

19.8

Elongation (%)

95

7.4

94

5.5

93

6.8

97

4.5

Wet Set (%)

18.5

1.3

16.6

2.6

19.1

0.2

16.4

0.6

50% Compression Set

10.1

1.1

10.2

1.1

10.4

0.9

9.1

0.8

50% HACS

28.0

0.2

26.6

1.2

25.7

0.7

24.3

1.6

1.4.1.3 FTC Results The FTC graph shown in Figure 1.41 illustrates an approximate 14% decrease in initial FTC values and continues throughout the remaining FTC cycles. This improvement in FTC represents foam articles that will shrink less immediately after demoulding when cooling of the trapped gas begins, allowing more time to complete the post crushing

66

Dimensional Stabilising Additives for Flexible Polyurethane Foams procedure or potentially eliminating the procedure all together. Several foams were prepared with each surfactant and allowed to cool for 24 hours in an uncrushed state, to confirm the relationship between FTC and foam shrinkage. Foams prepared with Dabco DC2517 and Dabco DC5258 showed significantly less shrinkage, approximately 1015% less, as compared to the control and Dabco DC2585. This is readily comparable to the differences in FTC values shown in Figure 1.42.

Figure 1.42 FTC Graphs for MDI Dimensional Stability Additives

1.4.2 Low Emissions Dimensional Stability Additives in MDI The control catalyst combination for the MDI seating is a standard commercial catalyst package. The balanced non-fugitive catalyst package Dabco NE1060/Dabco NE200 and the low emission catalyst package Dabco NE1060/Dabco BL-11 use levels were set in order to match performance to the commercial catalyst control package for the MDI formulations. The silicone surfactant control was Dabco 2525. Cell opening non-fugitive catalysts were not needed for this application based on the reported FTC values discussed later in this chapter. An MDI formulation, with a density of 40 kg/m3 and an MDI formulation, with a density of 55 kg/m3 were used in the MDI automotive study with the formulations shown in Table 1.35. For rapid demoulding of non-fugitive MDI systems, it is recommended that Dabco NE1060 be used at 0.5 to 2.0 pphp, with the optimum level at 1.0 to 1.5 pphp with 0.3 to 0.7 pphp of Dabco NE200. For low emission systems the authors recommend Dabco NE1060 be used at 1.0 to 2.0 pphp, with the optimum level at 1.0 to1.6 pphp with 0.03 to 0.1 pphp of Dabco BL-11.

67

Advances in Urethane Science and Technology

Table 1.35 Flexible moulded MDI formulations Formulation Identification

XXV

XXVI

XXVII

XXVIII

XXIX

XXX

XXXI

XXXII

XXXIII

XXXIV

Components

pphp

pphp

pphp

pphp

pphp

pphp

pphp

pphp

pphp

pphp

Polyol C

100

100

100

100

100

100

100

100

100

100

Polyol D

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

Water

3.5

3.5

3.5

3.5

3.5

3.5

3.5

3.5

3.5

3.5

Control Catalyst Package

0.70

-

-

-

-

0.70

-

-

-

-

Dabco BL-11

-

0.08

-

0.05

-

-

0.08

-

0.05

-

Polycat 15

-

0.30

0.30

-

-

-

0.30

0.30

-

-

Dabco NE1060

-

1.00

0.55

1.50

1.20

-

1.00

0.55

1.50

1.20

Dabco NE200

-

-

0.60

-

0.40

-

-

0.60

-

0.40

Dabco DC2525

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

Crosslinker

0.60

0.60

0.60

0.60

0.60

0.60

0.60

0.60

0.60

0.60

90

90

90

90

90

105

105

105

105

105

Modified MDI Index

1.4.2.1 Low Emissions Moulded Foam Machine Standard Physical Properties For complete comparison of these new additive technologies that provide low volatility, no amine emissions and no fogging, all moulded formulations were run on a Krauss-Maffei machine. Several foams were produced for each catalyst combination and formulation to obtain physical property pads and FTC pads, shown in Tables 1.36 to 1.41. Tables 1.36 and 1.37 provide the physical property comparison for the ~46 kg/m3 MDI flexible moulded formulations XXV-XXXIV. The data clearly demonstrates that physical properties are maintained, and in several cases improved, compared to the control formulations, depending on the formulation and the use of experimental cell-opening catalysts. For example, the data in Table 1.36 illustrates that at an index of 90, all physical properties are comparable to the control formulation. When the index is increased to 105, ball rebound and airflow are improved, with a slight decrease in indentation force deflection (IFD) properties when the non-fugitive catalyst (XXXIV) is compared to the control formulation (XXVI). When Dabco NE1060/Dabco NE200 is used (XXIX), the physical properties are equal to the control at an index of 90. Improved Japanese wet sets (see Section 1.3.1.2f) and slightly decreased IFD values are observed at an index of 105 compared to the control formulation (XXX).

68

Dimensional Stabilising Additives for Flexible Polyurethane Foams Table 1.36 Physical properties for 46 kg/m3 MDI moulded formulations at an index of 90 Physical Property

XX V

XXVI

XXVII

XXVIII

XXIX

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

159

-

159

-

14 9

-

15 5

-

15 5

-

65%

459

-

450

-

43 6

-

45 0

-

44 8

-

25% Return

127

-

126

-

11 9

-

12 3

-

12 3

-

51

0.6

51

0.6

51

0.6

51

0.6

51

1.2

Airflow (SLM)

Ball Rebound (%)

22.4

7.6

24.4

6.5

29

7.4

34

7. 9

17

5.0

Density (kg/m3)

46

3.4

44

1.0

43

2.4

46

3.7

45

0.9

Tensile (kPa)

129

10.6

298

14.4

125

6.7

133

5.9

127

9.9

Tear (N/m)

212

12

201

16

197

12

217

16

205

20

Elongation (%)

114

9.9

118

6.1

112

4.6

120

1.3

118

9.3

Wet Set (%)

13

0.6

16

1.2

10

0.5

13

0. 4

13

1.3

50% Compression Set (%)

9

1.6

10

1.0

9

1.6

10

0.5

10

0.8

50% Humid Aged Compression Set (%)

13

0.5

13

2.0

13

1.5

12

1. 3

14

1.0

Table 1.37 Physical properties for 46 kg/m3 MDI moulded formulations at an index of 105 Physical Property

XX X

XXXI

XXXII

XXXIII

XXXIV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

275

-

260

-

264

-

261

-

251

-

65%

768

-

731

-

739

-

738

-

71 6

-

25% Return

209

-

197

-

201

-

198

-

190

-

Ball Rebound (%)

43

0.6

45

1.2

45

0.0

45

1. 0

47

1.2

Airflow (SLM)

40.8

8.5

31.4

7.1

38

27.2

33

9.9

46

8.5

Density (kg/m3)

46

1.1

47

4.4

45

0.6

46

3.2

44

1.3

Tensile (kPa)

176

9.8

409

19.2

173

6.8

168

6.3

169

11.7

Tear (N/m

263

8

252

4

25 6

12

279

16

244

12

Elongation (%)

105

4.2

109

4.1

102

6.7

105

4.3

103

6.2

Wet Set (%)

15

0.8

15

0.4

11

0.2

9

0. 7

17

0.2

50% Compression Set (%)

8

0.6

9

1.3

8

0.3

10

0.9

9

1.0

50% Humid Aged Compression Set (%)

13

2.0

13

1.8

12

1.0

14

1. 7

14

1.0

69

Advances in Urethane Science and Technology Table 1.38 Physical properties for 55 kg/m3 MDI moulded formulations at an index of 90 Physical Property

XXV

XXVI

XXVII

XXVIII

XXIX

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

235

-

238

-

218

-

26 0

-

23 5

-

65%

637

-

636

-

59 7

-

731

-

64 1

-

25% Return

188

-

190

-

17 4

-

197

-

18 7

-

Ball Rebound (%)

46

0.6

44

1.0

46

0.6

45

1.2

42

0.6

Airflow (SLM)

21

4.0

35

11.9

35

5.4

31.4

7.1

27

12.5

Density (kg/m3)

53

3.4

50

3.6

50

3.5

47

4. 4

52

1.9

Tensile (kPa)

148

24.7

164

8.0

161

3.2

149

19.2

162

11.0

Tear (N/m

207

12

220

12

197

4

252

4

20 5

12

Elongation (%)

117

6.0

116

7.9

120

1.7

109

4.1

118

4.1

Wet Set (%)

11

1.4

13

0.9

15

1.3

15

0. 4

10

0.3

50% Compression Set (%)

8

0.9

7

1.5

9

1. 0

9

1. 3

9

0.4

50% Humid Aged Compression Set (%)

11

1.5

14

2.1

13

1.3

13

1.8

12

0.9

Table 1.39 Physical properties for 55 kg/m3 MDI moulded formulations at an index of 105 Physical Property

XXX

XXXII

XXXI

XXXIII

XXXIV

ILD (N)

AVG

SD

AVG

SD

AVG

SD

AVG

SD

AVG

SD

25%

398

-

38 0

-

38 8

-

38 3

-

381

-

65%

1066

-

1029

-

1039

-

1042

-

1032

-

25% Return

303

-

288

-

295

-

289

-

289

-

Ball Rebound (%)

46

1.0

43

0.6

44

0.6

43

0.6

45

1.2

Airflow (SLM)

54

17.3

62.9

21.8

48

15.0

67

9.3

48

4.5

Density (kg/m3)

52

3.7

52

1.2

54

0.7

54

0. 4

55

0.5

Tensile (kPa)

217

17.0

218

7.0

207

12.9

217

6.6

223

10.5

Tear (N/m)

299

12

283

24

307

24

29 9

16

315

16

Elongation (%)

105

4.2

108

1.4

102

1.1

111

4.1

109

2.1

Wet Set (%)

14

1.4

12

0.7

13

0.7

11

0. 6

15

0.9

50% Compression Set (%)

6

1.1

8

0.8

8

0. 6

9

1. 1

9

0.4

50% Humid Aged Compression Set (%)

14

1.9

14

0.8

13

0.9

12

1. 2

14

0.8

70

Dimensional Stabilising Additives for Flexible Polyurethane Foams Tables 1.38 and 1.39 provide the physical property comparison for the ~55 kg/m3 MDI flexible moulded formulations XXV-XXXIV. The data clearly demonstrates that physical properties are maintained, and in several cases improved, compared to the control formulations, depending on the formulation. For example, the data in Table 1.38 illustrates that all physical properties are matched to the control formulation (XXV) at an index of 105 with the exception of a slightly lower ball rebound for the full non-fugitive catalyst package. At an index of 105, all physical properties for the experimental non-fugitive catalysts formulations were determined to be similar to the control (XXX). When Dabco NE1060/BL-11 is used as a low emission catalyst package (XXXIII), all physical properties at the index of 105 are comparable to the control foam. At an index of 105 (XXXIII), airflow, elongation and Japanese wet sets and HACS are improved over the control (XXX) and 50% dry compression sets are slightly elevated.

1.4.2.2 FTC Tables 1.40 and 1.41 show numerical values which are shown graphically in Figures 1.43 and 1.44. FTC values are dramatically improved versus the control (XXV), shown

Table 1.40 FTC values for MDI formulations at an index of 90 FTC (N/323cm2) 46 kg/m3

FTC (N/323cm2) 55 kg/m3

XXV

52

76

XXVI

14

24

XXVII

35

63

XXVIII

8

13

XXIX

4

12

Formulation

Table 1.41 FTC values for MDI formulations at an index of 105 FTC (N/323cm2) 46 kg/m3

FTC (N/323cm2) 55 kg/m3

XXX

2.6

4.5

XXXI

1.3

1.9

XXXII

1.9

3.2

XXXIII

1.9

3.2

XXXIV

1.3

1.9

Formulation

71

Advances in Urethane Science and Technology

Figure 1.43 FTC at an Index of 90 for MDI Formulations

Figure 1.44 FTC at an Index of 105 for MDI Formulations

in Figure 1.43, for the foams with an index of 90. FTC values for all of the formulations at an index of 105 are very similar to the control (error is ± 6.4 N/323cm2), as illustrated in Figure 1.44.

72

Dimensional Stabilising Additives for Flexible Polyurethane Foams

1.4.2.3 MDI Flexible Moulded Foam Review The non-fugitive catalyst package Dabco NE1060/Dabco NE200 and the low emission catalyst combination Dabco NE1060/Dabco BL-11 closely match the entire reactivity profile demonstrated by the standard control catalysts typically used in MDI flexible foam. Additionally, the Dabco NE1060/Dabco NE200 combination demonstrates a higher selectivity towards the blowing reaction as compared to a standard 2.5:1 ratio of Dabco 33LV to Dabco BL-11. This higher blowing selectivity can help improve cell opening and produce a more dimensionally stable polyurethane article. These non-fugitive catalysts also provide a reduction in the force necessary to crush freshly demoulded foam without adversely affecting the remaining physical properties of the foam in MDI formulations over a range of formulation indices.

1.5 TDI Flexible Slabstock Low Emission Additives Handmix rate of rise comparison tests were performed using the same equipment and process as described for TDI and MDI hand mix foam. The formulations used are listed in Table 1.42 and the rate of rise data reported in Table 1.43. Dabco BLV with Dabco T-9 was used as the industry standard catalyst control for the water blown flexible slabstock foam formulations. The silicone surfactant used in the reported flexible slabstock formulations was DC5160. All water blown formulations were made at densities of 23 kg/m3.

Table 1.42 Flexible slabstock formulations Formulation Identification

XXXV

XXXVI

XXXVII

Components

pphp

pphp

pphp

Voranol 3512

100

100

100

Water

4.60

4.60

4.60

Dabco BLV

0.12

—

—

Dabco NE500

—

0.06

0.16

Dabco NE600

—

0.12

—

Dabco T-9

0.25

0.25

0.26

Dabco DC5160

0.90

0.90

0.90

TDI (100 index)

57.4

57.4

57.4

73

Advances in Urethane Science and Technology

Table 1.43 Hand mix free rise reactivity comparison using flexible slabstock formulations Formulation

XXXV

XXXVI

XXXVII

Cream (Seconds)

13

12

13

Top of Cup (Seconds)

25

24

24

String Gel (Seconds)

67

67

64

Full Rise (Seconds)

112

1 11

109

1.5.1 Reactivity Formulations from Table 1.42 were used in rate of rise catalyst comparisons and the results are summarised in Table 1.43. The data suggests there is no significant change in reactivity when comparing the control formulation (XXXV) to the experimental formulations.

1.5.2 Standard Physical Properties Table 1.44 lists the physical property data for the flexible slabstock formulations. The experimental non-fugitive catalyst Dabco NE500 can be used with a slight increase in Dabco T-9 catalyst level (shown in Table 1.42, formulation XXXVII). Formulation XXXVII compares very well to the control (XXXV). Formulation XXXVI illustrates the use of Dabco NE500 and Dabco NE600 balanced non-fugitive catalyst combination, to match a Dabco BLV control. In this case, all physical properties were also matched, with the exception of slightly decreased airflow and improved ILD properties. Overall, either the non-fugitive catalyst combination Dabco NE500/Dabco NE600 or the use of Dabco NE500 with increased Dabco T-9 can replace Dabco BLV in all water blown flexible slabstock formulation.

1.5.3 TDI Flexible Slabstock Foam Review Dabco NE600 in combination with Dabco NE500 or Dabco NE600 as a sole amine catalyst used with increased levels of T-9 provides for equal or improved performance compared to Dabco BLV in all water-blown flexible slabstock foam. In all cases, these catalyst combinations can completely replace Dabco BLV and eliminate amine emissions from the flexible slabstock foam.

74

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Table 1.44 Physical properties for flexible slabstock foam Physical Property

XXXV

XXXVI

XXXVII

ILD (N)

AVG

SD

AVG

SD

AVG

SD

25%

198

-

211

-

200

-

65%

401

-

421

-

401

-

25% Return

126

-

134

-

128

-

Ball Rebound (%)

41

0.1

40

0.1

40

0.1

Airflow (SLM)

108

5.3

72

6.0

94

6.8

Density (kg/m3)

22

0.3

22

0.4

22

0.2

Tensile (kPa)

71

0.6

73

0.5

75

0.5

Tear (N/m)

200

-

200

-

200

-

Elongation (%)

97

7.5

106

6.7

104

6.1

90% Compression Set (%)

6

0.5

7

0.5

6

0.5

1.6 Foam Model Tool Discussions Details of the TDI and MDI foam model systems have been previously published [2]. The models require the use of mono-functional reactants that are quantitatively analysed to correlate structure-activity relationships for various classes of catalysts. A realistic thermal profile is produced through the imposition of an external exotherm. Urethane, urea, allophanate and biuret reaction products are quantified by liquid chromatographic analysis of quenched reaction samples. The models effectively account for such nonideal conditions as reactant depletion at variable rates, temperature and concentrationdependent catalyst activity, and catalyst selectivity as a function of isocyanate distribution.

1.6.1 TDI and MDI Moulded Foam Model The information below highlights the features characteristic of the TDI and the MDI flexible moulded models. Table 1.45 illustrates the conversion of the formulations from Tables 1.12-1.15 into a TDI model system. Table 1.46 illustrates the conversion of the formulations from Table 1.35 into an MDI model system. Since the TDI and MDI automotive polyols are highly ethylene oxide (EO) tipped the reactive hydroxyl group can be represented by a primary alcohol. The electronic effect of

75

Advances in Urethane Science and Technology

Table 1.45 TDI flexible moulded foam model system Formulation

Mass (g)

Polyether Polyols

100

Model Compounds

Mass (g)

Diethylene glycol methyl ether CH3OCH2CH2OCH2CH2OH

10.43

Diglyme CH3OCH2CH2OCH2CH2OCH3

95.89

Diethanolamine (water free basis)a

1.74

Dibutylamine

2.14

Water

4.20

H2O

4.20

Surfactant

No surfactant required for model

Catalyst Package

0.35

Catalyst Package

TDI 80 Index

105

40 mole% PhNCO 60 mole% o-TolylNCO

a

0.35 Varied

Diethanolamine hydroxyls included with diethyleneglycol methyl ether mass.

Table 1.46 MDI flexible moulded foam model system Formulation Polyether Polyol

Mass (g) 100

Model Compounds

Mass (g)

Diethylene glycol methyl ether CH3OCH2CH2OCH2CH2OH

7.51

Diglyme CH3OCH2CH2OCH2CH2OCH3

93.54

Diisopropanolamine (water free)a

0.85

Dibutylamine

0.81

Water

2.90

H2O

2.90

Dabco DC-2525

1.00

No surfactant required for model

Catalyst Package

1.60

Catalyst Package

Mondur MR Index a

95

80 mole% PhNCO 20 mole% o-TolylNCO

1.60 Varied

Diisopropanolamine hydroxyls included with diethyleneglycol methyl ether mass.

76

Dimensional Stabilising Additives for Flexible Polyurethane Foams the ether two carbons away from the alcohol is accounted for by using an ether alcohol. Thus both stearic and electronic effects on the reaction kinetics can be accurately reflected in the model compound. The unreactive polyether backbone of the polyol can be represented by the dimethyl ether of diethylene glycol. Dipropylene glycol dimethyl ether would be more representative of the polyether backbone of a typical EO-tipped propylene oxide (PO) polyol, but the rate and selectivity measurements are minimally affected by the choice of model polyol backbone. Note the sum of the masses of the model alcohol and the diglyme slightly exceeds 100 grams, the total polyol mass, since the hydroxyls from the alkanolamine crosslinkers are included in the model alcohol mass. Nonetheless, the concentration of reactive groups is comparable to that in the actual foam. Water and catalysts are the same for both formulation and model. The model system does not require a surfactant however, since a foam is not actually produced. TDI 80 and Mondur MR are represented by 40:60 mole% and 80:20 mole% mixtures of phenyl and orthotolyl isocyanate, respectively. Phenyl isocyanate represents the 4-position in 2,4-TDI or 4,4´-MDI, and tolyl isocyanate represents the more sterically hindered 2- and 6-positions in the 2,4- and 2,6-TDI, or the internal rings of MDI oligomers. It is important to include both isocyanate types because using only the more reactive phenyl isocyanate significantly overestimates the reactivity of the system. Detailed procedures for an individual model run can be found in the literature [14]. For these runs a masterbatch mixture containing the alcohol, ether, water, catalyst and the 3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatch mixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouple well, septum, and jacketed mechanical stirrer with gas inlet and a septum. For the TDI model the flask was placed under a slow argon purge in a sand bath capable of raising the internal temperature from 60 °C to 120 °C in 4 minutes in the absence of any reactions. The isocyanate was added via a syringe when the internal temperature reached 60 °C. For the MDI runs the flask was wrapped with insulation but not otherwise heated externally, and the isocyanate was added at 25 °C. Samples were then withdrawn via a syringe at 30 second intervals for eight minutes. The samples were quenched with dibutylamine and analysed by liquid chromatography to determine the yields of the urea and urethane reaction products as well as the unreacted isocyanates, analysed as the dibutyl ureas. The flexible moulded chemical foam model provides a detailed look at the performance of the industry standard catalyst package Dabco 33LV/Dabco BL-11 compared to the non-fugitive catalyst packages Dabco NE1060/Dabco NE200 and Dabco NE1060/XFN1085. The foam results presented previously can be explained in terms of the overall catalyst performance. The catalysts were characterised with a model system because urea, urethane and isocyanate can be quantified as a function of time, so this approach provides the highest level of detail. However, a model system is only relevant to the

77

Advances in Urethane Science and Technology extent that it accurately reproduces the environment in an actual foam. The mixing in the model systems is less energetic, so this may influence the relative rates of catalyst hydrolysis versus reaction with isocyanate. However, if catalysts are compared under a consistent set of conditions, the relative performance differences should be meaningful. The most convenient way to compare catalyst performance is to use blow to gel selectivity. Selectivity as a function of time (t), is defined next. Normalised blowing rate = (% yield of ureas at time, t) / (limiting urea yield) Normalised gelling rate = (% yield of urethanes at time, t) / (limiting urethane yield) Blow to gel selectivity = (Normalised blowing rate) / (Normalised gelling rate) Catalyst selectivity is defined as the ratio of the normalised amount of blowing (urea formation) to the normalised amount of gelling (urethane formation). A selectivity of 1.0 means that the normalised amounts of blowing and gelling are equal at that point in the reaction. A selectivity substantially below 1.0, for example about 0.3, is indicative of a strong gelling catalyst. A selectivity greater than 1.0 is indicative of a blowing catalyst. However, it has been shown that a catalyst with a blow-to-gel selectivity greater than 0.8 can still serve to balance a strong gelling catalyst such as TEDA, and a good quality foam can be produced [14]. Note the limiting urea and urethane yields are simply the molar equivalents of water and alcohol, respectively, from Tables 1.45 and 1.46. Selectivities are plotted as a function of isocyanate conversion in Figures 1.45 and 1.46.

1.6.2 TDI Flexible Slabstock Foam Model Details on the development and use of foam model systems for flexible moulded foam have been published [14]. In this section the features characteristic of slabstock foams will be highlighted. Table 1.47 illustrates the conversion of the formulations from Table 1.42 into a model system. Since the Voranol 3010 polyol is fully PO tipped, the reactive hydroxyl group can be represented by a secondary alcohol. The unreactive polyether backbone of the polyol can be represented by the dimethyl ethers of dipropylene glycol and diethylene glycol. The diglyme represents the wt% EO incorporation into the polyol. Note the sum of the masses of the model alcohol and the two dimethyl ethers equals 100 grams, the total polyol mass. This ensures that the concentration of reactive groups is comparable to that in the actual foam. Water and catalysts are the same for both formulation and model.

78

Dimensional Stabilising Additives for Flexible Polyurethane Foams

Figure 1.45 TDI Flexible Foam Model Selectivity versus NCO Conversion

Figure 1.46 MDI Flexible Foam Model Selectivity versus NCO Conversion

79

Advances in Urethane Science and Technology

Table 1.47 Slabstock foam model system Formulation Voranol 3010

Mass (g) 100

Model Compounds 1-Methoxy-2-propanol CH3OCH2CH(CH3)OH

9.01

Dipropyleneglycol dimethyl ether (CH3O(CH3)CHCH2)2O

82.99

Diglyme CH3OCH2CH2OCH2CH2OCH3 Water

2.00-6.00

Mass (g)

H 2O

8.0 2.00-6.00

Dabco DC5160

0.75

No surfactant required for model

Dabco BLV

0.20

Dabco BLV

0.20

Dabco T-9

0.50

Dabco T-10

0.50

TDI 80 Index

90, 105, 120 40 mole% PhNCO 60 mole% o-Tolyl NCO

Varied

Once again the model system does not require a surfactant, since a foam is not actually produced. As noted previously, TDI 80 is represented by a 40:60 mole% mixture of phenyl and ortho-tolyl isocyanate. Detailed procedures for an individual model run can be found in the literature [14]. For these runs a masterbatch mixture containing the alcohol, ethers, water, Dabco BLV catalyst and the 3,3´-dimethylbiphenyl internal standard was prepared in advance. The masterbatch mixture was charged to a 50 cm3 roundbottom flask equipped with a glass thermocouple well, septum, and a jacketed mechanical stirrer with gas inlet. The flask was placed under a slow argon purge in a sand bath capable of raising the internal temperature from 50 °C to 120 °C in 4 minutes in the absence of any reactions. Isocyanate was added via syringe when the internal temperature reached 50 °C, and the T-9 was injected from a microliter syringe immediately afterwards. Samples were then withdrawn via the syringe at 30 second intervals for eight minutes. The samples were quenched with dibutylamine and analysed by liquid chromatography to give yields of the urea and urethane reaction products as well as the unreacted isocyanates, analysed as the dibutyl ureas. The chemical flexible slabstock foam model provides a detailed look at the performance of the industry standard catalyst Dabco BLV compared to the non-fugitive catalysts Dabco NE500 and the Dabco NE500/Dabco NE600 package, keeping the Dabco T-9 level constant. Selectivity and conversion can be calculated as described previously. Note that

80

Dimensional Stabilising Additives for Flexible Polyurethane Foams the limiting urea and urethane yields are simply the molar equivalents of water and alcohol, respectively, from Table 1.47. Dabco 33LV is considered to be a strong gelling catalyst, but under slabstock conditions it is actually more of a blowing catalyst, with an initial selectivity near 2. This reflects the difficulty of catalysing the secondary alcohol-isocyanate reaction in competition with the water-isocyanate reaction. Tertiary amine catalysts are highly sensitive to stearic hindrance near the reaction site. Secondary alcohols actually become less reactive than water, even in the presence of a strong gelling catalyst. This is why tin catalysts are critical in flexible slabstock formulations because they are less sensitive to stearically hindered reactions. Dabco BLV contains some bisdimethylaminoethyl ether (Dabco BL11), which is a very strong blowing catalyst. Thus the initial selectivity is even higher, near 4. The non-fugitive catalysts give selectivities comparable to that of Dabco BLV. Note that the non-fugitive catalyst packages are more blowing selective than both the TDI and MDI controls. Modern polyol technology tolerates and in some cases even benefits from higher initial blowing selectivity because of the reduced levels of monol and diol relative to triol. Thus network formation is more efficient, and correspondingly less gelling is required to produce a superior urethane network. Higher blowing selectivity can promote improved cell opening, which is now more critical in the face of improved network formation. In a sense the control formulations are over gelled, but older formulations could tolerate high gelling due to weaker urethane networks and the frequent use of Dabco BL-11, which is a potent blowing catalyst. Although the selectivity curves are different, the catalyst packages are still rate matched. Rate matching in foam is usually accomplished by comparing rise profiles, which are largely determined by the blowing rate. Thus the non-fugitive catalyst packages provide blowing comparable to the controls, but provide a lower extent of urethane formation early in the foaming process.

1.7 Conclusions Incorporation of these newly developed additives in the production of polyurethane foam results in polyurethane articles that give dimensional stability, low emissions and wider processing latitude. These dimensional stability additives provide a significant reduction in the force necessary to crush freshly demoulded foam without adversely affecting the physical properties of the foam. As increasing demands are placed on the foam producers to meet specific comfort and durability requirements, physical properties and environmental concerns, the need for these additives will continue to be a key criterion for polyurethane manufacturers.

81

Advances in Urethane Science and Technology

Acknowledgments The authors would like to thank and acknowledge the efforts and support of the following people at Air Products and Chemicals, Inc.: Mark A Eckert, Steven E Robbins for physical property testing of foams generated in this study, and Ilean S Ruhe for preparation and editing of this work. We would also like to thank Air Products and Chemicals, Inc., for support of this work and permission to publish it.

References 1.

J. D. Tobias, G. D. Andrew, Presented at the SPI Polyurethanes Expo ‘98, Dallas, TX, 1998, p.445.

2.

M. L. Listemann, A. C. L. Savoca and A. L. Wressell, Journal of Cellular Plastics, 1992, 28, 4, 360.

3.

R. G. Petrella and S. A. Kushner, Presented at the SPI Polyurethanes ‘90 Conference, Orlando, FL, 1990, p.186.

4.

G. Burkhart and M. Klincke, Presented at the SPI Polyurethanes ‘95 Conference, Chicago, IL, 1995, p.297.

5.

Flexible Polyurethane Foam, The Dow Chemical Company.

6.

ASTM D3574-95 Standard Test Methods for Flexible Cellular Materials - Slab, Bonded, and Molded Urethanes Foams.

7.

L. J. Gibson and M. F. Ashby, Cellular Solids, Structure and Properties, Pergamon Press, Oxford, 1988, Chapters 5 and 6.

8.

K. D. Cavender, Presented at the SPI, Magic of Polyurethane Conference, Reno, NV, 1985, p.314.

9.

SAE J1756 Test Procedure to Determine the Fogging Characteristics of Interior Automotive Materials, 1994.

10. M. S. Vratsanos, Presented at the SPI, Polyurethanes ’92 Conference, New Orleans, MS, 1992, p.248. 11. K. F. Mansfield, Air Products & Chemicals, Inc., unpublished results.

82

Dimensional Stabilising Additives for Flexible Polyurethane Foams 12. R. G. Petrella and J. D. Tobias, Presented at the SPI Polyurethanes ‘88 Conference, Philadelphia, PA, 1988, p.28. 13. D. G Battice and W. I. Lopes, Presented at the SPI Polyurethanes: Exploring New Horizons Conference, Toronto, Canada, 1986, p.145. 14. M. L. Listemann, K. R. Lassila, K. E. Minnich and A. C. L. Savoca, inventors; Air Products and Chemicals, Inc., assignee; US Patent 5,508,314, 1996.

83

Advances in Urethane Science and Technology

84

2

Demands on Surfactants in Polyurethane Foam Production with Liquid Carbon Dioxide Blowing Andreas Weier and Georg Burkhart

2.1 History of Polyurethane Foams The history of polyurethane foams and of polyurethanes in general is not that long for a plastic material. It has been characterised by one special feature: the consistent changes of the industry. When Otto Bayer started his research on this new class of plastic material built from the addition reaction of isocyanates and hydroxyl containing materials in the years 1936/1937 he envisioned a plastic material which could be tailored to many different applications by the broad variety of acidic hydrogen containing compounds already available. But there were also setbacks to this concept. The first major drawback was the fact that the necessary diisocyanates were not available on an industrial scale and large scale synthesis of these compounds seemed to be a rather high hurdle. A second major point was an obstacle that Bayer and his group faced again and again. When they tried to make the polyurethanes they envisioned in solid form, like films, etc., they always ended up with gas bubbles in their plastics. That was due to the inefficiency of the methods available to synthesise pure water free material. After encountering this effect for nearly four years Bayer made a bold move in line with the concept ‘use what you cannot prevent’ – in this case the formation of gas bubbles [1, 2]. Thus Bayer’s group started to make polyurethane foams. Although these new lightweight porous materials were envisioned as supportive materials as well as insulative materials in the base patents [3, 4] their market development remained slow. This can be seen by the fact that even in 1952 polyisocyanates, mainly toluene diisocyanate (TDI), were available worldwide in quantities of less than 100 tonnes. After this rather hesitant start of polyurethane history and the first major switch from solid materials to porous foamed plastics, the industry has been characterised by significant changes in concept and the resulting industrial application of these switches. The first major switch resulted from the basic research enabling the technical production of soft polyurethane foams in the early 1950s. This technology was mostly focused on polyester polyols as raw materials. About five years later polyether polyols entered the

85

Advances in Urethane Science and Technology market as yet another major change within the polyurethane foam industry. Since the early 1950s the foaming machines for the industrial production of polyurethane foams had been discontinuous machines. With the use of polyalkyleneglycols as polyols, the importance of continuous foam production grew significantly, for both performance and commercial reasons. In the early 1950s, the necessary stabilisers for the production of polyurethane foams had been silicone oils, with the major suppliers being Bayer, Dow Corning, General Electric and ICI. In the late 1950s the picture changed with the broad application of the continuous process, first with prepolymers, then the one-shot-process. For this type of polyurethane foam production a different class of stabilisers was needed: the new silicone-polyether copolymers. The larger variability of the available polyetherpolyols laid the foundation for a significant broadening of the performance range of polyurethanes – and all this with lower costs than before. The growth of polyurethanes after this chemical change was fast [5]. Even in 1960 more than 45,000 tons of flexible polyurethane foam were produced. Besides its major application as insulation materials and comfort products in furniture applications, new usage areas have been continuously identified and tailor-made polyurethane materials have been supplied. Nowadays the range covers very high density contour parts as well as very low density packaging foams, flexible foams as well as semi-rigid and rigid foams, thermoplastic polyurethanes as well as integral skin foams or reaction injection moulded (RIM) materials [6]. In the beginning of the 1990s an estimated 500,000 people had jobs in the polyurethane industry, either within production of raw materials or in the conversion to final products [7].

2.1.1 Environmental Concerns in Relation to Flexible Foam Density By definition the production of a foam depends on the formation and stabilisation of gas bubbles in a liquid. That is true even for polyurethane foam when the additional curing of the liquid results in an elastic or rigid solid material. One of the central questions is how the gas bubbles are generated and how long this generation takes. The basic chemical gas forming process in polyurethane foaming is the very exothermic reaction of water and isocyanate, resulting in the generation of carbon dioxide, urea and heat. This method as the sole method to generate the gas bubbles of a polyurethane foam is limited to the production of a rather small range of density/hardness combinations. The achievable minimum density is limited by the tolerable heat in a foam bun before scorching or even self-ignition occurs. Increasing water levels lead to the generation of increasing amounts of carbon dioxide and therefore to lower densities. Increasing water levels will also result in higher hardness as more urea is generated. During reaction the urea appears in a variety of associated forms before it will ultimately separate as a solid in flexible slabstock polyether foams [8, 9, 10]. 86

Demands on Surfactants in Polyurethane Foam Production… During the history of polyurethane flexible slabstock foam production the accessible density/ hardness range was continuously broadened with a wide variety of techniques. The most obvious one was the use of a physical blowing agent to generate the gas volume necessary to expand or generate gas bubbles without generating additional heat during the process. Due to its inert character and its low biological toxicity, CFC-11 (Freon) was the liquid blowing agent of choice for many years in Europe, North America and Asia. That changed after recognition of the fact that CFC-11 had both a high global warming potential (GWP) and a high ozone depletion potential (ODP). The result was the inception of the Montreal Protocol and subsequently the exploration of a broad variety of alternative blowing agents (ABA), including hydrochlorofluorocarbons (HCFC), acetone, cyclopentane and methylene chloride. With the use of ABA, both the available densities and the available foam hardnesses were lowered in comparison to the purely water blown foams, as in these cases gas volume was produced without generation of urea. Softening of foam was also achieved chemically, either with the use of an additive to change the urea morphology [13] or with the use of crosslinkers in combination with a low isocyanate index [14]. A few years later machinery modifications enabled foamers to produce low density foams by the use of increased water levels while reducing traditional blowing agent levels. This reaction preferentially utilises the reaction generating carbon dioxide as a blowing agent. When the water in the formulation is increased, the foam bun temperature can become too high for safe processing. This problem has been tackled by forced cooling equipment generating an air exchange in the foam bun directly after production and thereby cooling the bun to prevent scorch or fire problems. There is a variety of technologies for the forced cooling of foams, including Envirocure [13, 14], Rapid Cure [15], Reeves Brothers [16] and lateral cooling [17]. For the production of very soft foam grades some physical blowing agent is still needed. Also there are always some safety concerns over possible equipment malfunctions when making these hot foam grades. A safer way to control flexible foam density and hardness is the variation of production pressure. Making high as well as low pressures available in the foam production environment can give access to higher density hard grades as well as to low density foams without the use of physical blowing agents. The two commercial processes addressing this approach are Variable Pressure Foaming (VPF) and Foam One. As foam production is done in a closed chamber, volatile materials can be trapped rather efficiently. This creates the possibility of a clean and safe, as well as ABA free foam production. A setback is the high capital investment cost, especially for a VPF unit. Therefore this equipment only becomes cost effective in very high volume production plants. Review of the many production techniques reveals that the production of a broad range of flexible foam qualities with reduction of environmental hazards (ODP as well as GWP) is possible by numerous options. All of these current technologies offer benefits, but provide no clear economical solution for total ABA elimination.

87

Advances in Urethane Science and Technology Within all the blowing agents available today, carbon dioxide probably offers the best combination of zero ODP, low GWP and low price. As it is abundant, it doesn’t have to be generated. For capturing and transporting only a minor amount of energy is needed. Merely the point of emission is changed. On top of all that it is also non-flammable, which minimises safety concerns in the foam plant. Therefore many attempts have been made to produce polyurethane foams with carbon dioxide that is not generated by a chemical reaction during the foaming process. The difficulty is that carbon dioxide is a gas at room temperature under atmospheric pressure. The only way to utilise carbon dioxide not generated chemically is to mix the carbon dioxide under pressure or in a liquid state with the chemical components of the foam formulation. Then the pressure is subsequently released causing the frothing or preexpansion. Carbon dioxide as a non-reactive blowing agent in the frothing technique has been suggested for quite some time already [18, 19]. The pressurised reaction mixture is ejected at atmospheric pressure causing a turbulent vaporisation of the blowing agent. This allows the manufacture of a foam with reduced density, but the cell structure is of very inconsistent quality due to irregular shaped and oversized cells or bubbles being present.

2.2 Current Liquid Carbon Dioxide Technologies for Flexible Slabstock Polyether Foam Production 2.2.1 Machinery The situation changed when Cannon introduced liquid carbon dioxide foaming technology for the industrial production of polyurethane flexible foams with its CarDio process. This process was introduced to the polyurethane foam production industry in 1994. Today foam manufacturers worldwide consider liquid carbon dioxide as a blowing agent and make this process available for the industry. The major point of this technology is the control of the frothing that occurs as soon as the pressurised liquid carbon dioxide as part of the foam formulation leaves the pressurised equipment and the liquid foam mixture is released into atmospheric pressure. The controlled release of the mixture into the comparatively low atmospheric pressure conditions is essential to all the liquid carbon dioxide foaming techniques. In the CarDio process this is accomplished with the specifically designed lay down device [20, 21] which is called a ‘Gate Bar’ (see Figure 2.1). The gate bar essentially consists of a metal bar with a pressure drop slot connecting the outside with a feeding tube inside the bar. Through a liquid mix injection point formulation components with the pressurised liquid carbon dioxide enter the gate bar.

88

Demands on Surfactants in Polyurethane Foam Production…

Figure 2.1 Cannon Gate Bar Principle

The frothing mixture flows along the frothing cavity and through an outlet aperture acting as a pressure drop zone onto a substrate, e.g., the moving belt of the foam machine. A variation of the original gate bar design was introduced with the CarDio 2000 featuring a pressure adjustable lay down device [22]. It allows adjustment of the slot width in response to the output and the liquid carbon dioxide percentage (see Figure 2.2).

Figure 2.2 CarDio Froth laydown

89

Advances in Urethane Science and Technology Another patented technology for liquid carbon dioxide foaming is the NovaFlex process developed by Hennecke and Bayer [23]. The NovaFlex process uses a creamer dispensing the froth on to the machine. It focuses on the same task of pressure reduction as the CarDio gate bar, except dispensing the froth occurs at one single spot on the belt and is not distributed over nearly the whole belt width. The same is true with the third version of liquid carbon dioxide machinery, the Beamech CO-2 equipment. Obviously, both the NovaFlex as well as the Beamech CO-2 process use pressure reduction devices different from the gate bar of the CarDio process as they are said to show a different response to the use of fillers or generally solid particles within the foam formulations. One modification of the original NovaFlex process is the MultiStream configuration. It allows the use of liquid carbon dioxide with a range of polyol types in all different ratios by utilising one liquid carbon dioxide addition point in one high pressure polyol stream which is then combined with additional polyols by pass streams [24, 25].

2.2.2 The Foaming Process 2.2.2.1 A Comparison of Rise Profiles A comparison of the foaming profile found in a standard flexible slabstock foam production with the one seen during liquid carbon dioxide foaming clearly shows a large difference in the rise profiles, especially at the lay down (Figure 2.3).

Figure 2.3 Foaming profiles of standard flexible foam production and liquid carbon dioxide foaming.

90

Demands on Surfactants in Polyurethane Foam Production… In standard foam production, a mixture of liquid materials is dispensed and the gas bubbles within the foam structure are only generated by a rather slow chemical reaction. In the case of liquid carbon dioxide foaming the mixture is already dispensed in a two-phase process. No matter which particular pressure drop system is utilised, the pressure drop is very fast due to the large flow speeds in the system. The fast pressure drop between the mixing head and the dispensing compresses this time span for bubble formation to fractions of a second. As all the gas bubbles resulting in foam cells of the final foam have to be generated in this short instant, clearly the nucleating powers of the formulation ingredients are challenged severely. This is especially true when looking at the foam stabilisers used as they are the main formulation ingredients determining the nucleation.

2.2.2.2 The Nucleation of Gas Bubbles As early as 1969, Kanner and Decker [26] showed by photomicrography that self-nucleation is essentially absent in a polyurethane foam system. Their results indicate that bubbles are introduced by the process of mixing and that the presence of a silicone surfactant increases the volume of air bubbles introduced during the mixing. Nucleation in this context means the formation of gas bubble cores on which a bubble might grow above its critical bubble radius. The critical bubble radius is the smallest bubble radius for which the addition of more gas will result in a net decrease in free energy. A bubble which is smaller than the critical radius will not spontaneously grow in size because the addition of more gas results in a net increase in free energy. Obviously a decrease in the surface tension of a foaming mixture by the silicone surfactant would decrease the energy needed to generate bubbles of the necessary critical radius. Even under conditions of a supersaturation pressure of 2.03 MPa carbon dioxide in a liquid with a surface tension of 0.025 N/m and a temperature of 25 °C, the critical bubble radius is 2.4 x 10-5 cm. It has been proven that silicone surfactant can aid nucleation by lowering surface tension [27]. However the surface tension lowering of a silicone surfactant might just be one effect helping to induce new bubble formation by liquid carbon dioxide. The surfactant is entirely soluble in the liquid and therefore cannot act as a heterogeneous nucleating agent. Besides that other surfactants like fluorocarbons can also lower the surface tension but still don’t act like a silicone surfactant does in polyurethane foam production. A possible explanation of this difference might be related to the exceptional solubility of oxygen and nitrogen in silicones as demonstrated by Arkles [28, 29]. It is well known in the polyurethane industry that insufficient air loading of raw materials like the polyols results in coarse cell structures even in the conventional foaming process. Obviously a sufficient amount of nitrogen and/or oxygen is needed for good nucleation. As silicones show a high permeability and solubility for these gases, it is reasonable to assume that

91

Advances in Urethane Science and Technology the ease of formation of nuclei for gas bubbles is higher in silicone-like surroundings provided along the siloxane backbone of a silicone surfactant. An effect not accessible to Kanner and Decker, and therefore not a subject of their experimental work, was the high nucleation demand found in the new liquid carbon dioxide foaming technologies. Their finding that bubbles are introduced during the process of mixing by the shear energy of the mixer might reflect the observation that the amount of nuclei generated during mixing has always been sufficient for a standard polyurethane foaming process. The rather large timescale between mixing and appearance of gas bubbles large enough to be visually detectable obviously gave carbon dioxide molecules sufficient time to diffuse to a nucleation site as they were only generated slowly within the foam formulation. With the new liquid carbon dioxide foaming technologies, a significantly higher number of nucleation sites is needed for the carbon dioxide to find a gas bubble or a gas bubble core during the short time frame of the pressure decrease in the pressure reduction area within the equipment used. Whereas the former trend in some polyurethane (PU) markets has been to sacrifice more and more of a surfactant’s nucleating power to gain higher and higher activity, liquid carbon dioxide foaming has led to a new description of the needed performance profile. This is recognising the fact that quite a number of the very high active surfactants used at the time of the introduction of the liquid carbon dioxide processes resulted in unacceptable foam structures. Irrespective of the liquid carbon dioxide system, the high activity surfactants generated coarse foams. That is in contrast to the findings of Kanner and Decker in standard foam systems. Obviously under these conditions the mixing energy of the machine system has not been high enough to generate sufficient nuclei or gas bubbles to guarantee a regular fine celled foam structure. Obviously under these conditions the main quality issue of a surfactant is its nucleation power, and not the stabilisation activity. Not surprisingly a well established medium active surfactant was used during the development of the first commercial liquid carbon dioxide process, the CarDio process. It provided a good balance between nucleation and processing as well as activity. Therefore this type of surfactant was still the number one choice for processes even four years after the introduction of commercial liquid carbon dioxide foaming. Still, due to the high demands in nucleation, the broad variety of foam grades and the flame retardant (FR) demands of some markets a desire for optimising the different surfactant performance aspects could clearly be seen. It also provided a focus of interest for the flexible slab polyurethane community for quite a few years.

2.2.2.3 Emulsification of Raw Materials In addition to nucleation, there is one further objective to be fulfilled or at least supported by the silicones used. That is the emulsification of the raw materials used in the foam 92

Demands on Surfactants in Polyurethane Foam Production… formulation. An obvious example for this performance aspect is the chemical reaction of water and isocyanate, i.e., the formation of carbon dioxide and urea. The different polarities of isocyanate and water should normally prevent their reaction with each other in the short time frame of a PU foam reaction and even the use of polyols as the major formulation component will not always help to overcome the tendency for phase separation. Even the polyols used in some special foam types can give rise to a number of problems. One of the challenges for the emulsification power of silicone surfactants in flexible slabstock foam production for example is the combination of a standard polyol with a hypersoft polyol. The consistency of the foaming results and the cell structure obtained can be severely influenced by the choice of silicone surfactant used. It can be easily rationalised that the rise profile and physical foam properties like tensile strength and elongation will be influenced by tendencies for phase separation after the mixing and dispensing of the formulation ingredients.

2.2.2.4 Stabilisation of the Foam During Foam Rise Probably the most obvious task of any silicone surfactant in flexible polyurethane foam production is to stabilise the foam during the foam rise. If the surfactant used is contaminated or is in too low a concentration, settling of the foam or even foam collapse will occur. Two possible reasons for the foam instability during the rise time are discussed in the literature, and lead to different theories on the roles of silicone surfactants used. The first explanation of the role of silicone surfactants in flexible polyurethane foam is based on the fact that as soon as the volume fraction of the gas bubbles exceeds 74%, the spherical bubbles will distort into multi-sided polyhedrals. This means that cell windows with Gibb’s plateau borders are formed (a polyhedral foam consists of cell windows and struts. Another term for struts is Gibb’s plateau borders). A pressure difference due to capillary pressure will cause liquid in the cell window to drain into the struts since the pressure inside the plateau borders is lower than that in the cell windows. Without adding silicone surfactants this drainage rate will be very fast, so that film rupture and bubble coalescence occur rapidly. It has been shown that due to the surface tension gradient generated by a silicone surfactant the cell window drainage rate is lower [30]. Therefore different surface structures do have different effects on cell window drainage. That has been said to result in a distribution of different cell window thicknesses at the time of cell opening [31]. The critical factor in this picture is the surface elasticity. It is known to be caused by a surface tension gradient along the cell window upon expansion and it will retard the cell window drainage rate. This effect is called Gibbs-Marangoni effect and is reviewed by Scriven and Sternling [32]. Owen and co-workers also studied the dynamic surface tension of a series of surfactant solutions and stressed the importance of surface elasticity [33]. Similar work has been performed by Zhang and co-workers [34].

93

Advances in Urethane Science and Technology A different view point of the mechanism of flexible polyether polyurethane foam stabilisation by silicone-based surfactants has been published by Rossmy and co-workers [8, 9, 10]. The authors investigated the urea precipitation phenomenon in polyurethane foams in greater detail. By adding a defoaming agent to a foam formulation they showed that urea precipitation turned a clear foam mix opaque. They also noted that the cell rupture always occurred just after urea precipitation. In light of this observation the hypothesis is that precipitation of polyurea destabilises the foam mix and leads to cell opening. The surfactant then aids in stabilising the foam by incorporating the urea precipitate in the foam matrix, adding integrity to the foam. Today it is not doubted any more that there really are urea domain structures in polyurethane foams and they have been subject to intensive research [35, 36].

2.2.2.5 Control of Blow Off Another major point of consideration is obviously the processing latitude of silicone surfactants. This phrase describes the variability of catalyst concentration without running into either closed, dead foam or observing first signs of foam instability indicated by splits in the foam. Not only will a different silicone surfactant result in a different processing window, e.g., amount of stannous octoate variation between the two extremes, but also the position of the processing window might be different. Figure 2.4 shows the remaining fragments of a cell window after cell opening as found in the fully cured foam. It is very obvious that the opening of a flexible PU foam is not a

Figure 2.4 Opening of a cell window

94

Demands on Surfactants in Polyurethane Foam Production… pressure induced rupture of a foam bubble. Otherwise there would be no explanation why the cell opening process clearly starts at a number of locations around the periphery of the cell window. Obviously weak points in the cell window offer starting points for the opening of the liquid film. Then drainage occurs, as can be seen by the thin strings remaining between the opened areas in the cell window. The actual timing of this opening process is critical as it determines the physical properties of the foam and the foam economics. As the viscosity of the liquid material is rising very sharply due to crosslinking at the end of the foaming profile, there is only a very short time frame in which the material is crosslinked enough to support the foam without a high settling once the gas bubbles have opened and no internal gas pressure is supporting the foam anymore. On the other hand a further delay in gas bubble opening would lead to an incomplete recession of the material into the foam struts so that the foam will be showing a rather low air flow or even shrinkage.

2.2.3 Additional Tasks of Silicone Surfactants in Flexible Slabstock Foam Production 2.2.3.1 Influence on burn performance of the foam As soon as the foam is produced and cured the silicone surfactants used are not needed or desired any more. One exception might be the production of hydrophilic or hydrophobic foams where a surface active material like a silicone will in some cases have a recognisable influence. But with standard foams as well as with liquid carbon dioxide blown foams, additional performance aspects besides the stabilisation during the foam rise might be important for surfactant selection. These criteria therefore have to be addressed as well and deserve some consideration as they might also affect the number of key additives a foam operation has to handle. An effect clearly seen in the final foam is the undesirable effect the silicone surfactant has on the burn performance of the resulting foam. Especially in liquid carbon dioxide foaming, suitable conventional (or non-FR) surfactants have been available right from the beginning. But the development of optimised universal or non-FR-surfactants for liquid carbon dioxide foaming took considerable effort and time. To understand the effect of the silicone surfactant on burning it is interesting to have a look at a theory for flame spread development. The simplified illustration in Figure 2.5 shows that the heat decomposes the organic material at the surface during an endothermic procedure. Pyrolysis products are created and because of an exothermic reaction with oxygen at the boundary layer of the flame, the formation of even higher reactive decomposition products takes place. These aggressive radicals are responsible for an accelerated degradation of the polymeric surface. As long as the result of the energy is positive, there is thermal feedback to the endothermic process

95

Advances in Urethane Science and Technology

Figure 2.5 General assumption about flame spread

at the surface and the combustion keeps going. It is believed that the formation of hydrogen radicals and hydroxyl radicals from the organic material are the key factors initiating and supporting the combustion phenomena. The working mechanisms of FR address these phenomena. The effect of a silicone surfactant on the burning behaviour of the polyurethane foam can be quite significant even though they normally represent less than 1% of the plastic material. It is due to the fact that in any case the decomposition of a polyurethane foam starts at the surface. Because of the surface activity of the foam stabilisers it is easy to rationalise their enrichment on the surface and it is the surface that is the most influential part of the polymer regarding flame spread development. This point can be further highlighted if we look at the influence of surfactants on the FR-performance of a polyurethane foam with 1 weight percent of surfactant externally applied (Figure 2.6). To characterise the burn performance in this case, a horizontal burn test was used and the burn length was measured. Surfactants A, B, C and D are rigid foam surfactants with different structural parameters in the silicone as well as in the polyether chains. It can be seen in Figure 2.6 that all the different types of silicone surfactants used do have a negative influence on the burn performance of the foam. The reason for this behaviour can be mainly attributed to the structure of the siloxane backbone within the siliconepolyethercopolymers [37]. To characterise the effect of the polydimethylsiloxane (PDMS) chain used as the backbone of such a siliconepolyethercopolymer, interesting tests have been made by external application of pure PDMS in different weights as well as the use of molecules of different average chain length (Figure 2.7).

96

Demands on Surfactants in Polyurethane Foam Production…

Figure 2.6 Influence of surfactants on FR-performance - external application of 1% surfactant

Figure 2.7 Effect of 1% PDMS - applied externally

Not only do increasing amounts of externally applied PDMS lead to increasing burn length in, but more astonishingly, this effect is obviously correlated to the average chain length of the applied siloxane. It can be seen that with longer chain length of the PDMS, the detrimental effect on the FR-performance of the foam increases. This effect far out balances the pure weight effect that would explain why a higher amount of PDMS leads to higher burn length value due to the larger amount of materials supporting the decomposition. To address the question of whether the different volatility of the different PDMS chains is the main reason for this effect, an additional test with silica was carried

97

Advances in Urethane Science and Technology out. It is easy to expect that solid materials, especially if inorganic, will have a positive effect on the FR-performance of the foam and that they will lead to a decreasing burn length in a burn test. Depending on the chemical nature they could be either expected to be a heat sink, a material to catch free radicals, to char the burn front of the foam or to liberate non-gaseous products to dilute the oxygen at the burn front. Astonishingly, silica, a chemically inert material, under these conditions increased the burn length of the foam (Figure 2.8). The assumption here is that on the surface of the silica-particles either catalytic effects on the heat initiated oxidation processes take place or the silica particles themselves act like a wick in a candle, providing a means to transport melted organic material from the surface and increase the diffusion flame front. If the formation of silica-particles on the burn front is indeed the major detrimental effect of the silicone containing polymer, it should be helpful to decrease the average length of any unmodified PDMS chain part between any two modifying groups. Although in practice this is also an efficient tool to improve the burn performance of the resulting foam, it negatively affects the nucleation power of the silicone surfactant. Therefore this way of addressing FR-performance behaviour is detrimental to the nucleation efficiency of any surfactant for liquid carbon dioxide foaming [38]. That is the main reason why conventional (non-FR) silicones for the liquid carbon dioxide processes were available before their optimised universal or even pure FR-counterparts.

Figure 2.8 Addition of solid material

98

Demands on Surfactants in Polyurethane Foam Production…

2.2.4 Chemistry of a Silicone Surfactant in Flexible Slabstock Foam Production As outlined previously the main reason for the interest in optimised surfactants for the liquid carbon dioxide processes lies in their role as nucleation promoters. The silicone surfactants can be viewed as PDMS-polyether-copolymers which are mainly based on a combination of just three structural units: the methyl substituted siloxane backbone as well as a sophisticated ratio and arrangement of ethylene oxide and/or propylene oxide forming the attached polyethers and, in some cases, additional modifications. A typical structure of a silicone surfactant is shown in Figure 2.9.

Figure 2.9 Building blocks of a surfactant

The molecules generally have a siloxane backbone formed by dimethylsiloxane units, substituted methylsiloxane units and some endgroups. Polyether groups and/or additional modifications can be attached to the siloxane backbone. Not only are a wide variety of those additional modifications possible but also the polyether groups can vary in quite a number of ways. Although they mostly consist of ethylene oxide and propylene oxide units they obviously can be different in chain length, i.e., their molecular weight. They can also have the different monomers distributed along the polyether chain in an either random or a blockwise fashion as well as with alternating blocks. The polarity can be either evenly or unevenly distributed. A more polar side could be either close to the attachment point to the siloxane or far away from it. As this is true for each of the polyethers used and flexible

99

Advances in Urethane Science and Technology slabstock surfactants generally have more than one type of polyether attached, this gives rise to virtually millions of different combinations for surfactant molecules. Also, in most foaming applications the pendant polyethers have to be unreactive to isocyanates so that they do not act like crosslinkers which would lead to tight or shrinking foam. In addition there are two general ways to link the polyethers to the siloxane chains (see Figure 2.9). This opens the basis for two separate product lines, each having its special advantages over the other.

2.2.4.1 Si-C-Surfactants Versus Si-O-C-Surfactants The Si-O-C products resulting from the reaction of chlorosiloxanes and hydroxyl groups of polyethers provide an extremely good processing and superior consistency combined with being hydrolytically stable under the water-amine conditions found in polyurethane foaming. The Si-C products offer a more beneficial access to high activity and lead to an easier production of flame retardant foams. These products are derived from the addition of an Si-H functionality at the siloxanes to a double bond, a process called hydrosilylation. Examples of how this variety of structural parameters affects the development of silicone surfactants especially for liquid carbon dioxide blown foams, are many [39, 40, 41, 42] and this latest drive in surfactant development provides good examples of the important performance issues and how they can be addressed.

Figure 2.10 Synthesis of Si-C and Si-O-C linked siloxane-polyether-copolymers

100

Demands on Surfactants in Polyurethane Foam Production…

2.2.5 A Surfactant Development Example In surfactant developments for liquid carbon dioxide PU foaming, the described complexity of structural parameters in a surfactant led to the desire to characterise nucleation performance of a surfactant without the need to judge the final, cured foam. One important reason is the fact that a coarse or even irregular cell structure in the final foam does not necessarily indicate the reason for such a ‘bad’ cell structure. It could arise from a deficiency in nucleation performance so that the number of gas bubbles formed at the beginning of the foaming reaction was not sufficient (lack of nucleation). Another possible scenario could be that the initial number of bubbles formed was rather high, but many of the gas bubbles were lost during the foam rise due to gas bubble coalescence. That would mean that the stabilising power of the surfactant would be insufficient and had to be improved. One of the first published attempts to get an indication of the nucleation performance of a silicone surfactant was the froth test [39, 40, 41]. In this test polyol and surfactant are stirred in a reproducible fashion and the resulting froth density as well as the time needed for the foam to collapse is measured. After some experience with industrial surfactants used in the market these products with known performance in industrial liquid carbon dioxide foaming were subjected to the froth test. No direct correlation could be found between the obtained data and the performance of the products on industrial machines (Figure 2.11).

Figure 2.11 Performance of surfactants (Tegostab B 8228 and B 8220) in the froth test Comp.: competitive surfactant; EP-H-18: experimental product

101

Advances in Urethane Science and Technology As can be seen from Figure 2.11, both products Tegostab B 8228 and Tegostab B 8220 show nearly identical performance in this mixing test although industrial experience has shown that Tegostab B 8220 results in significantly finer cell structure than Tegostab B 8228, even if both of them yield a very regular cell structure. Even more confusing is the picture resulting from the comparison of a widely used surfactant and an experimental product called EP-H-18 by Burkhart and co-authors [42]. Both of the products again resulted in identical values in the froth stability test, but differed significantly on an industrial production machine. Whereas the competitive material resulted in a coarse cell structure, the experimental product yielded a fine and regular foam.

A (30 s)

B (30 s)

C (30 s)

A (150 s)

B (150 s)

C (150 s)

bad

good

better

Figure 2.12 Video images of foam formulations with commercial surfactants of different performance in liquid carbon dioxide processes

102

Demands on Surfactants in Polyurethane Foam Production… Because of these results a video microscopy check on the homogeneity of the gas bubbles in the liquid mixture was performed. For this test the ingredients of a standard foam formulation were mixed and immediately poured onto a polished metal plate. The creaming of the foam was then filmed by a video camera with a 10x magnification. Thirty and 150 seconds after the stirring commenced, a video printout was made and the number of formed gas bubbles as well as their size homogeneity was characterised (see Figure 2.12). This procedure is referred to as the video test.

2.2.5.1 Possible Silicone Structures Not taking cyclic molecules into account, the general structures of industrial silicone surfactants for flexible slabstock foam production can be seen in Figure 2.13. The main building blocks of these materials are a PDMS backbone and attached polyethers based on ethylene oxide and propylene oxide addition products. The siloxane backbones can either be linear or branched and can have their polyether substituents attached in an either pendant or terminal location. These four general structures are outlined in Figure 2.13).

Figure 2.13 General structures of silicone surfactants (schematic depiction)

103

Advances in Urethane Science and Technology The simplest type of the possible structures is a linear A-B-A-copolymer with a straight siloxane backbone and polyether chains at the termini. This type of structure is described as ST (for straight, terminal). With a straight siloxane backbone there is also the possibility of attaching polyether groups as pendant chains. This combination is denoted SP (straight, pendant). If the siloxane is branched as well it could have polyether groups attached as terminal or as pendant groups (BT and BP). These four types of structures gave different results in the video imaging test (see Figure 2.14). The ST type of structure resulted in only a small number of gas bubbles being formed in the early stages of the reaction. Somewhat better was the combination of branched silicone with pendant polyether groups. The best video imaging results were reported with the branched siloxane, combined with terminal polyether groups followed closely by the SP type of molecule.

ST

SP

–

+

BT

BP

+(+)

o

Figure 2.14 Video imaging test, comparison of different silicone structures -, o, +, +(+) are a non numerical quality rating of degree of nucleation. - is poor and +(+) is very good

104

Demands on Surfactants in Polyurethane Foam Production… These results matched, more or less, with industrial experience. Whereas the ABA type of copolymer (ST-structure) is known to be a rather poor performing surfactant for polyurethane production, the SP and the BT type of structures are commonly used. The structure that performed best in the video imaging test (the BT type structure) is the standard structural type that can be found in commercial SiOC-PDMS copolymers obtained by the reaction of a chlorosubstituted PDMS with the hydroxyl group of polyether (see previously). Although these SiOC products (named after the silicone-oxygencarbon bond linking the PDMS and the polyether side chain) provide a very good combination of broad processing, good nucleation and consistent quality, their main disadvantage is their very recognisable negative effect on the burning behaviour of the produced polyurethane foam. Therefore these types of products are commonly denoted as conventional or non-FR-silicones. In contrast, the second best type of structure seemed to be the SP type. These structures are the fundamental means to build SiC-products which are used as universal surfactants, for example in North America. Although SP structures seemed to be the obvious choice to build FR or universal type of surfactants (having a less negative effect on FR-performance) they did not seem to show as good a nucleation as the BT type of molecules. The fourth structural type, BP products, no known experience or industrial application was recognised, and due to the combination of not extremely good nucleation (according to the video test) and the complicated synthesis necessary to make them, no reported work was undertaken with these types of molecules. Besides the type of branching in the siloxane backbone and the attachment points of the polyether side groups, another factor characterising the siloxanes is the number of unmodified PDMS groups for each of the modified methylsiloxane groups in the molecule. This ratio is often denoted as the P-value and is especially important in regards to the burn performance of the resulting foams (see Figure 2.10 and discussion above). As that is another variable that has to be taken into account when characterising a silicone surfactant for polyurethane foam production, this factor was screened for as well. As can be seen from the video test results, there seems to be a slight improvement in the nucleation efficiency with increasing P-value. However the drawback is that the higher P-values result in surfactants which will lead to a considerably negative effect on the burn performance of the resulting foam [43]. The summary of the data at this point seemed to indicate that the combination of good nucleation efficiency with the BT-structures and the advantage of high P-values seem to

105

Advances in Urethane Science and Technology

P-value Figure 2.15 Video imaging test; nucleation efficiency with increasing P-value

correlate with the fact that, especially based on SiOC type of chemistry, it is easier to build a non FR-surfactant structure than to build a universal or true FR-silicone surfactant for liquid carbon dioxide foaming. In the case of good FR-surfactants therefore, a compensation for the unavailability of high P-values had to be found in regard to nucleation efficiency of the molecule. Because of a suggestion in the literature [44] that the high solubility of carbon dioxide in certain aromatic solvents might be due to interaction between the molecules, the influence of aromatic structures in surfactant molecules was screened. Whereas the standard siloxane backbone in PU surfactants is normally consisting of a PDMS chain between the substituted silicon atoms, high temperature silicone resins are used with phenylmethylsiloxane groups within the chain. Although these types of molecules seemed to be viable candidates for testing, they did not seem to be attractive options for industrial use due to the complicated synthesis and the high costs associated with them. Another option to incorporate aromatic structures in organomodified silicones has been the use of styrene oxide as a monomer for the synthesis of the polyether side chain within the copolymer structure [45] (see Figure 2.16). Comparison of structural siliconepolyether analogues with and without styrene oxide incorporation in surfactant test molecules showed an indication of a positive trend in the video test regarding the nucleation efficiency of those aromatic group substituted silicone polyether copolymers. Surprisingly enough in a real foam test these molecules performed rather badly, resulting in a considerably coarser foam than the reference (see Figure 2.17).

106

Demands on Surfactants in Polyurethane Foam Production…

Figure 2.16 Aromatically substituted polyether

Surfactant without aromatic groups

with aromatic groups

Cell structure in foam test: 10 cells/cm

8 cells/cm

Figure 2.17 Comparison of standard and aromatically substituted siliconepolyether

That observation was rationalised with the following assumption. If indeed aromatic rings would increase the solubility of carbon dioxide then these types of structures would be more soluble in a liquid carbon dioxide containing environment. This would result in a decreased tendency to move to the surface of a such a liquid phase. The obvious result would be a decrease of surface activity, so the net result could very well be a higher nucleation efficiency combined with a lower activity during the rise time of the foam, leading to gas bubble coalescence.

107

Advances in Urethane Science and Technology If that is true, the polarity and therefore solubility of the molecule would have to be adjusted to lower values. As explained previously the increase of the P-value was not an option for the generation of a surfactant with good FR-performance. Therefore to further decrease the solubility of a surfactant in a formulation for flexible slabstock PU-foam, the polarity of the surfactant molecule was decreased by lowering the polarity of the polyether side chain with the use of higher alkylene oxides to have more non-polar groups along the polyether chain (Figure 2.18). This would lead to a higher carbon:oxygen ratio in the polyether side chain. Since it has been known for more than 30 years [46] that the polyether polarity affects the cell size distribution of flexible slabstock PU foams that is a very good method for fine-tuning surfactants. Foaming tests showed that the carbon:oxygen content had a significant impact on the overall activity of the surfactant, resulting in decreasing foam height with increasing carbon:oxygen ratio, especially when a carbon:oxygen ratio of about 2.6 was exceeded.

Figure 2.18 Influence of carbon to oxygen ratio in polyether side chains

The utilisation of the outlined principles and the continuous work with the many possible variables by all the additive suppliers again and again has brought new high performance surfactants, supporting the continuously changing demands in the innovative polyurethane industry and most probably will continue to do so. Therefore this industry has all the fascination that comes from the interaction of chemistry and physics, industrial application and theoretical interest, commercial importance and environmental awareness. So even more than 50 years after its beginning it will remain an area of new challenges and performance oriented efforts. 108

Demands on Surfactants in Polyurethane Foam Production…

References 1.

No inventor; I.G. Farben AG, assignee; German Patent 728,981, 1937.

2.

A. Höchtlen and W. Droste, inventors; Farbenfabriken Bayer AG, assignee; German Patent 913,474, 1941.

3.

P. Hoppe, inventor; Farbenfabriken Bayer, assignee; German Patent 851,851, 1948.

4.

W. Droste and A. Höchtlen, inventors; Farbenfabriken Bayer, assignee; German Patent 860,109, 1952.

5.

Chemical and Engineering News, 1961, 39, 11, 62.

6.

K. Moser, Kunststoffe, 1983, 73, 12, 764.

7.

F. K. Brochhagen in Handbook of Environmental Chemistry, Vol. 3, Part G, Ed., O. Hutzinger, 1991, Springer Verlag, Berlin, p.73.

8.

G. Rossmy, H. J. Kollmeier, W. Lidy, H. Schator and M. Wiemann, Journal of Cellular Plastics, 1981, 17, 6, 319.

9.

G. Rossmy, H. J. Kollmeier, W. Lidy, H. Schator and M. Wiemann, Journal of Cellular Plastics, 1977, 13, 1, 26.

10. G. Rossmy, W. Lidy, H. Schator, M. Wiemann and H J. Kollmeier, Journal of Cellular Plastics, 1977, 15, 5, 276. 11. G. Burkhart, H-H. Schlöns and V. Zellmer, inventors; Th. Goldschmidt AG, assignee; US Patent 5 132 333, 1992. 12. R. Ricciardi, Presented at the Polyurethane Foam Association conference in Scotsdale, AZ, 1996. 13. M. A. Ricciardi, D. J. Smudin, R. D. Wagner, M. Pcolinsky and J. E. Chaya, inventor; no assignee, US Patent 3,890,414, 1975. 14. M. A. Ricciardi and D. G. Dai, inventors; Crain Industries, Inc., assignee, US Patent 5,171,756, 1992. 15. H. Stone, inventor; PMC, Inc., assignee, US Patent 5,128,379, 1992. 16. A. A. Grizwold, inventor; Reeves Brothers, Inc., assignee; US Patent 4,537,912, 1985.

109

Advances in Urethane Science and Technology 17. J. L. Drye and G. C. Cavenaugh, inventors; Trinity American Corporation, assignee, US Patent 5,188,792, 1993. 18. P. Merriman, inventor; Dunlop Rubber Co., Ltd., assignee; US Patent 3,184,419, 1965. 19. E. N. Doyle and S. Carson, inventors; no assignee, US Patent 5,120,770,1992. 20. C. Fiorentini, M. Griffith and A. Charles, inventors; Krypton International SA, assignee; European Patent 0645226A2, 1995. 21. C. Fiorentini, M. Taverna and T. Griffiths, Journal of Cellular Polymers, 1994, 13, 5, 361. 22. CarDio Newsletter No.4, September 1997, Article no. 97037. 23. R. G. Eiben, Presented at Utech ’96, The Hague, The Netherlands, 1996, Paper No.31. 24. Innovations, 1998, October, 801, 15. 25. Novaflex Multistream, Hennecke leaflet PI125, October 1998. 26. B. Kanner, T. G. Decker and G. Thomas, Journal of Cellular Plastics, 1969, 5, 1, 32. 27. B. Kanner, W. G. Reid and I. H. Petersen, Industrial Engineering Chemistry, Product Research & Design, 1967, 6, 2, 88. 28. B. Arkles, Silanes & Silicones catalogue, Petrach Systems, Bartram Road, Bristol, PA, 19007, USA, 1987, p 87. 29. B. Arkles, Chemtech, 1983, 13, 542. 30. X. D. Zhang, C. W. Macosko, H. T. Davis, A. D. Nikolov and D. T. Wasan, Journal of Colloid and Interface Science, 1999, 215, 2, 270. 31. K. Yasunaga, R. A. Neff, X. D. Zhang and C. W. Macosko, Journal of Cellular Plastics, 1996, 32, 5, 427. 32. L. E. Scriven and C. V. Sternling, Nature, 1960, 187, 186. 33. M. J. Owen, T. C. Kendrick, B. M. Kingston and N. C. Lloyd, Journal of Colloid Interface Science, 1967, 24, 2, 141.

110

Demands on Surfactants in Polyurethane Foam Production… 34. X. D. Zhang, C. W. Macosko and H. T. Davis in Polymeric Foams, Ed., K. C. Khemani, Chapter 9, ACS Symposium Series 669, American Chemical Society, Washington DC, 1997. 35. J. P. Armistead, G. L. Wilkes and R. B. Turner, Journal of Applied Polymer Science, 1988, 35, 3, 601. 36. M. W. Creswick, K. D. Lee, R. B. Turner and L. M. Huber, Presented at the SPI 31st Annual Technical/Marketing Conference, Philadelphia, PA, 1988, p.11. 37. A. Weier, G. Burkhart, M. Klincke, Presented at Industrial Applications of Surfactants IV, Suffolk, UK, 1998, p.260. 38. G. Burkhart, R. Langenhagen and A. Weier, Presented at the Polyurethanes Expo ’98, Dallas, TX, 1998, p.129. 39. S. B. McVey, B. L. Hilker and L. F. Lawler, Presented at the Polyurethanes Foam Association conference, San Antonio, CA, 1995. 40. S. B. McVey, B. L. Hilker and L. F. Lawler, Presented at Utech 96, The Hague, Paper No.38. 41. G. Burkhart, V. Zellmer and R. Borgogelli, Presented at the SPI Polyurethanes Expo ’96, Las Vegas, NV, 1996, p.144. 42. G. Burkhart, R. Langenhagen and A. Weier, Presented at the Polyurethanes Expo ’98, Dallas, TX, 1998, p.129. 43. A. Weier, G. Burkhart and V. Zellmer, Presented at the SPI, Polyurethanes ’94 Conference, Boston, MA, 1994, p.202. 44. J. H. Hildebrand, J. M. Prausnitz and R. L. Scott, Regular and Related Solutions: the Solubility of Gases, Liquids and Solids, Van Nostrand Reinhold Co., New York, 1970. 45. No inventors; Th. Goldschmidt AG, assignee; German Patent 19, 726, 653, 1999. 46. R. J. Boudreau, Modern Plastics, 1967, 44, 5, 133.

111

Advances in Urethane Science and Technology

112

3

Polyurethane Processing: Recent Developments Max Taverna

This chapter highlights three aspects of manufacturing technology, which have recently brought significant benefits to the producers of automotive seats, reinforced parts and generic moulded items. These different manufacturing areas have been grouped here under a single title in order to cover a wider spectrum of reader’s interests and to illustrate the broad area of improvement and innovation that can still be found in this steady-growing segment of industry.

3.1 Industrial Solutions for the Production of Automotive Seats Using Polyurethane Multi-Component Formulations Cannon has developed complete systems for the manufacture of moulded polyurethane automotive seating elements made with varying combinations of several raw materials. The resultant cushions – although produced in a random sequence on the same moulding line - are characterised by mechanical properties tailored to the specific application of each part [1]. This section describes the components of this technology which are: •

dedicated multi-component, high-pressure metering machines with closed-loop control of both the output and pour pressures

•

dedicated mixing heads capable of processing six components (all with high-pressure recirculation), including low-viscosity toluene diisocyanate (TDI) and water-based additives

•

dedicated mould-handling systems for a flexible manufacturing concept based on just-in-time methods

3.1.1 Market Requirements A recent analysis (internal Cannon Report) of the automotive seat-manufacturing market segment, to identify the future needs and trends of this group highlighted a number of interesting considerations for a producer of polyurethane processing equipment.

113

Advances in Urethane Science and Technology First, the number of suppliers is getting smaller and smaller, due to the extensive merger and acquisition campaign run in recent years by the two largest players (Lear Co., and Johnson Controls). This concentration creates an opportunity for making strategic choices regarding the ‘make or buy’ decision with respect to manufacturing solutions. These large manufacturers have grown significantly thanks to their own successful development of both chemical formulations and moulding plants. Providing them with new machinery sometimes means accepting their request to manufacture a concept that was developed by their own engineering department. The second aspect concerns the growing demand for simpler, high-efficiency plants, with a reduced number of operators and a high degree of operating flexibility. It must be possible to produce several different parts on the same moulding line so that, in the case of a sudden change in production plans, complete projects can be switched from one production line to another within a very short time frame. This increased flexibility requires careful design of the metering equipment, mix heads and mould carriers because they must be able to perform very different tasks in sequence or by project. A specific request involves the potential to process multi-component formulations, where a wide range of foams can be produced on one machine. The third aspect, potentially in conflict with the previous concept, involves the development of families of formulations based on specific chemicals: all diphenylmethane diisocyanate (MDI), MDI/TDI in various percentages, all TDI and special polyols. This means dealing with very different demould times and moulding conditions, that render a generic ‘seat plant’, that was still so useable just a few years ago, obsolete. In this case, specific packages must be available, which conflicts with the concept of high flexibility expressed above.

3.1.2 Dedicated Solutions: Metering Equipment A high-pressure multi-component metering unit was designed for automotive seating producers, with a pump-dosing system capable of precisely dosing TDI and other lowviscosity components. Output adjustment on-the-fly and closed-loop control, easy to obtain with piston-driven metering units, were engineered for pump-driven machines as well. The number of main components (polyols and isocyanates) can be set, as necessary, since the design is modular: each dosing line includes a dedicated tank, its temperature control system, high- and low-pressure filters, recirculation valves, a dosing pump and motor and a portion of the control panel (see Figure 3.1). The modular design of this unit allows for easy addition of further components to a formulation. Storage, metering, temperature-time control modules and feed lines are assembled in modules and can be added as needed. Computerised process control easily integrates the new chemicals in the formulation.

114

Polyurethane Processing: Recent Developments

Figure 3.1 A high-pressure multi-component metering unit

Two methods were used to enable fast, precise changing of the output and to ensure closed-loop control of the machine, which is essential to guarantee continuous, accurate output of low-viscosity chemicals: •

via an inverter mounted on the pump motor

•

via a programmable step-by-step motor that changes the setting of the pump.

The output control on the closed-loop machine operates continuously, comparing the set output value of the inverters with the real output as measured by the volumetric flow transducers. The system enables pouring only when the parameter is within the limits set by the operator via the keyboard and when it is possible to change the pump speed in less than half a second. The unit also includes maintenance and alarm menus. Through the maintenance menu, it is possible to set limitations on the following parameters: •

number of shots

•

material consumption

•

working hours

115

Advances in Urethane Science and Technology Once these set values are reached, the corresponding warning message informs the operator to perform the required maintenance operation. The programmable stepping device is very precise - the pump can be set at 256 discrete positions - and can be easily programmed through the machine programme logic controller (PLC). Since the switching time between formulations can be as low as 0.6 of a second, multiple ratios can be set within the same pour program to produce multi-hardness and multi-density foams with repetitive results.

3.1.3 Dedicated Solutions: Mixing Heads Cannon has designed a new mixing head, capable of mixing six components, each with individual recirculation control. The new mixing head was specifically designed to meet the needs of the automotive seat producers who wished to mould TDI-based flexible foams with a maximum of flexibility in their formulation. Several pure ingredients are kept separate up to the point of injection and it is possible to operate each stream on demand, provide highpressure recirculation and avoid the contamination of components (see Figure 3.2) [1, 2].

Figure 3.2 The new Cannon Ax Head, specifically designed for multi-component polyurethane formulations. It can mix up to six chemicals all with a high-pressure recirculation feature. 116

Polyurethane Processing: Recent Developments Although this solution is currently available on the market, competitive mix heads are limited to four streams, their dimension and weight require heavy-duty pour robots while providing speed and pour pattern limitations during operation and they have complex pressure set-up and regulation procedures.

3.1.3.1 Multi-Component Operation The innovative aspect of this mixhead is that TDI - and eventually a TDI-compatible sixth component - are fed axially into the mixing chamber through the small piston that cleans the mixing chamber. The other four components, polyols, other catalysts, additives, flame retardants, are fed radially into the mixing chamber. A seal on the small piston provides a permanent separation between the polyol and TDI feeding areas so that any crossover of the low-viscosity components is avoided. The main advantage of the new mixing head is its reliability. It can operate for millions of shots because of the remote position of the TDI feeding area. TDI is fed in at a pressure of only 1 MPa, with perfect mixing efficiency (see Figures 3.3, 3.4 and 3.5).

Figure 3.3 Section of the mixing-chamber’s cleaning piston. Four chemicals are fed radially in the mixing chamber, while two are fed through a hole drilled in the cleaning piston.

117

Advances in Urethane Science and Technology

Figure 3.4 During the high-pressure recycle phase all six components flow in the grooves carved in the mixing chamber’s piston and return to their storage tanks.

Figure 3.5 When the injection signal is received, the mixing chamber’s piston retracts, the four components fed radially meet the two that are fed axially through the mixing chamber’s piston. The mixed blend reaches the discharge duct, which is positioned at an angle of 90° to the mixing chamber and leaves the head through the pouring hole.

118

Polyurethane Processing: Recent Developments Defining ‘Formulation #1’ as being composed of polyol #1 and isocyanate and ‘Formulation #2’ as polyol #2 and isocyanate, an example of a working sequence is as follows: 1. Pour ‘Formulation #1’: polyol #1 and isocyanate recycle through the mixing chamber’s piston grooves and are poured in the open mould 2. ‘Formulation #2’ remains dormant since polyol #2 continues to recirculate through a high-pressure nozzle positioned close to the mixhead 3. The formulation change is implemented by: • closing the mixing chamber’s piston (the self-cleaning piston is kept open) • opening the polyol #2 high-pressure nozzle and closing the polyol #1 nozzle • re-opening the mixing chamber’s piston with the new ‘Formulation #2’ The maximum response time, which occurs between two consecutive shots with different formulations, is 0.6 seconds.

3.1.3.2 Pour Pressure Control The pressure control of the component injectors, which is very important for ensuring the proper mixing of the different liquid streams, has been achieved via three different approaches: •

closed-loop control of the injection pressure via hydraulic servo valves and feedback control from the pressure gauges

•

open-loop control via hydraulic servo valves

•

fixed-position control via various hydraulic valves, preset at different values

The package supplied with the head to achieve pressure control (see Figure 3.6) includes: •

hydraulically-operated nozzles to be installed on the mixhead for the components that are fed radially; each nozzle has a double function: - selection of the component to be used in the specified formulation - control of the re-circulation/injection pressure during the pour

•

hydraulic unit for operation of nozzles

•

valves for control of the oil flow; each valve has three positions (off, injection pressure #1 and injection pressure #2)

119

Advances in Urethane Science and Technology

Figure 3.6 Scheme of the circuit used to guarantee a closed-loop control of the pouring pressure of each chemical.

•

high-performance recycle stream distributors with relevant pressure control to maintain non-specified components in high-pressure recirculation, ready to be injected within 0.6 seconds

•

one set of high-pressure flexible hoses to connect the proportional valves to the injector nozzles

•

appropriate controls

3.1.3.3 Variable Geometry Mixhead To ensure proper mixing conditions for a wide range of formulations - which can differ in chemical composition, viscosity, ratio and output - it is important to provide appropriate backpressure in the mixing area. This can be obtained with the adjustable geometry regulation of the mixing area: the mix head’s larger piston can be mechanically set to partially block the outlet of the mixing chamber when it is fully retracted. This occlusion increases the turbulence in the mixing area, causing more efficient mixing to be obtained.

120

Polyurethane Processing: Recent Developments

3.1.3.4 Advantages A very compact design: its outside dimensions of 40 x 20 cm and 22 kg weight translate into a high degree of manoeuvrability. •

The pour pattern can be defined very precisely, according to the design of the mould and the position of any inserts.

•

The pouring operation can be executed in an open mould without fear of collision with the upper mould half.

•

The robot carrying the head can achieve high acceleration and speeds without mechanically stressing the moving elements.

The large number of chemical components that can be handled simultaneously provides a high degree of flexibility in formulations, allowing for optimum use of the moulding line. Several different parts can be produced on the same moulding line and different types of foam can easily be produced in random sequence without forcing the operator to use pre-defined sequences of moulds.

3.1.4 Dedicated Solutions 3.1.4.1 Foaming Robots The basic features of a typical Cannon double-arm pour robot, normally supplied to carry two heads (see Figure 3.7), are: •

two-axes Cartesian robot with two arms moving independently of each other,

•

arm movements (along the ‘X’ and ‘Y’ axes) achieved via a rack-and-pinion system driven by electronic variable-speed motors,

•

racks fitted with position encoders to detect the relevant ‘X’ and ‘Y’ coordinates.

The Technical Specifications include: -

maximum speed for X axis: maximum speed for Y axis: maximum acceleration: maximum load on the arm: maximum ‘X’-axis stroke: maximum ‘Y’-axis stroke: minimum distance between 2 arms: available working area:

2.4 m/s 2.4 m/s 3.5 m/s2 120 kg 250 cm 115 cm 55 cm 307 x 115 cm

121

Advances in Urethane Science and Technology

Figure 3.7 A double-arm Cartesian robot, designed to perform multi-hardness PU foaming in open-mould pouring processes for automotive seat moulding.

3.1.4.2 Mould Carriers The concept of the dedicated flexible foam moulding line relies on a mould carrier having a simple solid structure, with a minimum number of mechanical parts, which require maintenance. It consists of two upper and two lower sections linked by means of a hooking system. The presence of two separate lower platens allows the operator to compensate for any difference in mould thickness without incurring any problem by mounting two moulds on the same carrier (see Figure 3.8). Insertion, centring and fixation of various moulds within the mould carrier are achieved by means of tensioning screws. Each mould carrier is usually provided with two pneumatic cylinders for opening and closing. Upon demand, the closing system can be actuated hydraulically or mechanically. Mould carrier tilting, for optimum evacuation of air during the filling phase, is accomplished with a cam system. Each mould carrier can have two tilting positions: either horizontal or frontally inclined by 20 degrees (see Figure 3.9). It is possible to programme different angles of opening and tilting to correspond with the type of mould currently being used. Mould identification – necessary in order to transmit the appropriate mould/formulation information to the dosing system - is achieved with a code reader. The electrical controls

122

Polyurethane Processing: Recent Developments

Figure 3.8 Independently-moving lower platens allow for the use of different thickness of moulds as a double mould carrier.

Figure 3.9 A different degree of tilting can be obtained for each mould carrier, to optimise the evacuation of air from the moulds during the filling/ polymerisation phase.

123

Advances in Urethane Science and Technology for the moulding line are interfaced to the metering system controls in order to coordinate all the working processes. The typical working sequence can be defined as follows: •

mould identification: the control system on the moulding line reads the code coming from the field

•

pour programme selection: inside the mixhead manipulator control unit, a schedule - assigning a specific pour programme to the different moulds - has to be loaded

•

pour programme execution: the manipulator control unit sends the input to the machine control unit; when the mould under consideration requires multi-hardness foaming, both mix heads are activated, while in the case of a single-hardness seat, only one mixhead is activated

Stroking of the Platens As stated previously, the lower half of the mould carrier is composed of two independent platens. They are fitted with a guiding system to ensure maximum parallelism between the platens. In order to ensure equal clamping pressure and complete closure of the moulds over the entire parting line, pneumatically inflated tubes stroke each platen. They are inflated after the mould carrier has been closed and deflated just before opening. The total stroke is a few centimetres. The clamping force afforded by this system is 12,700 kg, using a working pressure for the air bags of 0.3 MPa.

Mould Temperature Control System Individual thermoregulator units are used, one for each press. They are mounted on the rear of each mould carrier. This solution affords complete autonomy to each press, making it possible to control the temperature of the moulds prior to being placed on the production line. Subsequently, when a press is placed on the moulding line, it is ready to commence production. The thermoregulators work in a closed circuit; the relevant refilling has to be performed off-line. The control panel for each thermoregulator is mounted on the operator side for easier access.

124

Polyurethane Processing: Recent Developments Mould Carrier Exchange The mould carrier concept for fixing and centring is designed for quick mould carrier removal or exchange. The mould carrier can be removed as follows: • • •

disconnect the electrical plug (positioned on the operator side) disconnect the air by removing the relevant joint mounted on the mould carrier remove the complete mould carrier (with its thermoregulator) using a fork lift

The estimated time to perform this operation is less than five minutes.

Service Station A dedicated station where service operations on moulds and mould carriers can be performed, without disturbing the production cycle, is foreseen for each moulding line (see Figure 3.10).

Figure 3.10 All maintenance and setting operations on moulds and mould-carriers can be executed off-line, in a dedicated service station that performs all the line’s movements and operative functions.

125

Advances in Urethane Science and Technology The service station comes with one complete mould carrier on its freestanding frame. It is equipped with a dedicated thermoregulator for warming of moulds destined to be inserted on the moulding line, along with all the required safety fences/light barriers, controls, and connections for air, water and electrical power.

3.1.4.3 Transport Systems Cannon provides different mould-carrying systems, each customised to meet the customer’s needs. The most common solutions are conveyors and turning tables. In the last part of this section some innovative concepts, which provide a more compact layout, a minimised investment requirement and a maximum degree of flexibility in a changing manufacturing scenario, are presented.

Oval Conveyor The conveyor is made of a central steel frame, which is positioned on the floor. It is composed of a series of straight-line modules with a curved module at each end and comes complete with a number of carriages running on an oval steel track (also mounted on the floor). On each carriage, mainly composed of a rigid steel frame positioned on four pivoting wheels, a single lid mould-carrier is mounted. Four idle wheels, running along a central guide plate, maintain the system on its set path (see Figure 3.11). Continuous movement is achieved by means of one drive system being placed between every two carriages. Each driving system has two entrainment wheels running along the guide plate. The entrainment wheels – operated via an AC motor – are maintained in traction along the guide plate by a spring. An inverter is used to adjust the speed from 4 m to a maximum of 9 m/min, while the AC motor is fitted with a brake so that, in case of an emergency, the plant can stop within a few centimetres of travel. This system can easily be expanded to incorporate any new business obtained for that range of moulded products. Extensions are achieved by fitting an even number of new carriers to the line and extending the supporting frame accordingly.

Mould Carrying Systems: FlexiDrum Developed in the early 1980s as a revolutionary tool for the production of foamed refrigerator doors, the well-known concept of the rotary polymerisation system is now being proposed as a compact and simple mould-carrying system for the production of automotive seats (see Figure 3.12).

126

Polyurethane Processing: Recent Developments

Figure 3.11 A modular oval conveyor for flexible foam moulding. It can be easily extended, adding modules of track and pairs of mould carriers, when higher productivity is required.

Figure 3.12 Flexidrum, a compact moulding line for medium-low volume of automotive seat production. 1a Service station: demoulding, cleaning, insert positioning, release agent application; 1b Open mould view of 1a; 2 Foaming station, with robot on a platform; 3-6 Polymerisation stations; 7 Six arm rotating structure to support mould carriers; 8 Central collector of signals, air and warm water; 9 Supporting structure (and rotor, not shown)

127

Advances in Urethane Science and Technology Six or more mould carriers can be fixed on the surfaces of a wheel that rotates vertically – similar to a Ferris wheel, rather than horizontally, as with a merry-go-round. The opening and closing movements can be hydraulic or pneumatic. Mould conditioning can be accomplished using a rotating collector that feeds each mould carrier with water at the desired temperature. Manual service operations are performed on the lower mould half in the first station. Foaming is achieved in the subsequent station using a platform-mounted robot. Polymerisation takes place in the remaining four (or more) elevated stations as they are passed through a compact suction hood that removes the escaping fumes. Several advantages can be highlighted for this new version of mould carrying system: •

simplified mechanical construction that does not incorporate wheels which roll along the floor picking up and transporting scrap foam and/or dirt

•

optimised layout which will work even when space is limited

•

reduced power cost thanks to its vertical layout which minimises the suction area, requiring only one fan for extraction of the fumes

•

limited number of operators required

Carousel for In Situ Foaming A number of dedicated moulding lines have been designed for the production of in situ moulded foams. In situ moulding technology adds a delicate operation to the list of conventional operations (mould cleaning, release agent spraying, insert positioning, foaming and demoulding) that must be executed to mould a standard foamed item: the manual positioning of the textile container into which the foam will be dispensed. This is a delicate operation that requires some time, yet should not penalise the cycle. A practical solution consists of a carousel line with a row of service positions where the operators can work on moulds that have been temporarily taken off-line (see Figure 3.13). When the press leaves the curing area, it passes in front of the first free operator and is automatically disengaged by the dragging system. The moulds can be serviced, taking all the time required, then, when the textile inserts have been positioned, the carrier can be reinserted in the first available position in the line.

Multi-Hardness and Multi-Density Foams with Natural Carbon Dioxide The availability of multi-component mixheads with variable geometry opens the path to another interesting option: the addition of natural carbon dioxide for expansion of the

128

Polyurethane Processing: Recent Developments

Figure 3.13 Oval moulding line with 16 double mould carriers, designed for in situ foaming of textiles. 1 to 4 Service stations, where mould carriers are taken temporarily off-line for mould service; 5 Spare station; 6 Foaming station; A Pouring robot; B Metering equipment, on mezzanine; 7 to 16 Polymerisation stations; C Mould carrier changing station; D Suction hoods

foam with an environmentally friendly blowing agent and reduction of the foam density by as much as 20-25%. With this new mixhead, it is possible to use two different approaches to add natural carbon dioxide to the formulations: •

CannOxide [3-12] - a technology developed to meter natural carbon dioxide at the point of injection, at the desired percentage, into one of the polyol streams

•

EasyFroth [1, 2, 4, 11, 12, 13,14] for carbon dioxide – a technology that allows the premixing of given percentages of blowing agent in one of the two components, usually the isocyanate.

Both methods are currently in industrial production, each having operating and investment ‘pro’s and con’s’ that must be evaluated according to the production volumes and flexibility required. For example, a car seat with three different hardnesses can be produced using two different formulations for the hard and soft parts - (see areas ‘A’ and ‘B’ in Figure 3.14) - and incorporate some natural carbon dioxide to decrease the seat density below the thighs (area ‘C’).

129

Advances in Urethane Science and Technology

Figure 3.14 Triple hardness automotive seat A – Soft foam; B – Hard foam; C – Soft and lower-density foam, blown with natural carbon dioxide

3.2 ‘Foam & Film’ Technology - An Innovative Solution to Fully Automate the Manufacture of Automotive Sound Deadening Parts Until now, one of the major limitations in the polyurethane moulding process has been the necessity to interrupt the working sequence between each moulding, to remove foam scrap and apply release agent prior to foaming. The introduction of ‘Foam & Film’ technology makes manual intervention unnecessary, removing the one factor that has always been a major weakness when working with polyurethanes in comparison to other injected or extruded plastics [15-19]. The main idea behind this new approach consists of thermoforming a thermoplastic or polyurethane film as part of the moulding sequence. By using a vacuum effect, this film adheres perfectly and smoothly to the mould cavity, without any creases or wrinkle formation. The mould is equipped with a dedicated frame device, specifically designed to hold the film. A heating system ensures that the film reaches the desired temperature prior to the thermoforming phase and subsequent injection of polyurethane into the mould.

130

Polyurethane Processing: Recent Developments Several industrial applications have been found and a number of fully automated plants currently incorporate this technology. The manufacture of sound deadening parts for the automotive industry represents one of the most exciting applications for this innovative technology.

3.2.1 The Problem Mould cleaning is required because of the chemical nature of the polyurethane process. The objective in developing the ‘Foam & Film’ technology was to eliminate the tedious, time-consuming manual operations, which must be performed on the moulds as part of a discontinuous polyurethane foaming process. The need to properly vent the mould, to avoid air entrapment, often results in a thin flash of polyurethane being formed around the moulded part, which needs to be manually removed afterwards via a simple trimming operation. These manual operations involve both cleaning of the mould and spraying of the release agent. It is necessary to demould parts in the shortest possible time, when the foam is not yet fully polymerised, because the thin cross-section of the flash makes them very fragile during the first few minutes following the demould. Often small pieces of flash will break off when the part is extracted, falling back onto the mould. If they are not removed from the mould surface, they could mar the surface of the subsequent parts or they could allow rising foam to leak from the mould if left along the seal. More extensive cleaning operations are required every few shifts in order to remove deposits of release agent from the mould surface and there require production to be stopped. Various technologies have been applied to minimise the required downtime, but so far a valid solution has not been found. The adhesive force between polyurethane and metal (either aluminum or steel) requires a release agent to be sprayed onto the moulds every single cycle (or every few cycles) to enable easy removal of the part. In addition, various internal mould release technologies are available, but they do not apply to all formulations. As stated previously, these operations take a long time, are expensive and are typically executed manually and it is impossible to have a fully automated line without getting rid of them. To create a fully automated line, both the need to manually clean the mould and the need to spray release agents on it had to be eliminated.

3.2.2 The Approach to a Solution The new approach to polyurethane moulding is to introduce a thermoplastic or polyurethane film thermoforming process on one or both halves of the mould as part of

131

Advances in Urethane Science and Technology the overall sequence. By using a vacuum, this film adheres perfectly and smoothly to the mould cavity, without any creases or wrinkle formation. The system’s key concept is a specially designed and patented frame, integrated into the mould, which is specifically designed to hold the film in position during this thermoforming process. This frame can be either two or three dimensional, to better follow the shape and the cavities of the mould (see Figure 3.15).

Figure 3.15 ‘Foam and Film’ concept 1. Unrolling and cutting plastic film to required length; 2. Infrared (IR) heating of film on holding frame; 3. Thermoforming the film in the mould cavity; 4. The same process (1-3) is carried out in the other half of the mould at the same time; 5. Robotised foam deposition in mould; 6. Mould closed for expansion/polymerisation. Parts are removed ‘wrapped’ in the film.

The frame receives the film after it is unrolled by pinchers and cut dimensionally. It then moves it in front of an infrared lamp for several seconds to heat it to the required temperature. A control system ensures that the film reaches the desired temperature prior to the thermoforming phase. Special infrared heaters, Cannon’s MVL heaters, are used to ensure high efficiency and very low thermal inertia. Once the film is heated, the frame moves into the mould where the film, held in place along its four edges, is vacuum-formed onto it. The use of the frame during the vacuum forming prevents the film from folding or wrinkling (see Figure 3.16).

132

Polyurethane Processing: Recent Developments

Figure 3.16 A patented frame holds the releasing film in position so that when thermoformed over the mould, it is kept under tension and does not form wrinkles.

3.2.3 The Film There are two different types of ‘Foam & Film’ processes available, based on the different types of film used: adhesive and releasing. •

In the adhesive type, the film sticks to the part and is unloaded with it, granting an aesthetic finish and a waterproof covering (see Figure 3.17).

•

The releasing film, on the contrary, remains vacuum-formed to the mould for several shots (5-15, dependent upon the process and the materials utilised) and it is then replaced when it begins to wear.

Obviously, with the adhesive type, the film has a high adhesion coefficient with polyurethane, while in the releasing version the poorer the adhesion, the better. In both cases, the film prevents the mould from ever being put in contact with the polyurethane foam. That is why both cleaning and spraying are no longer necessary and all operations can be automated. 133

Advances in Urethane Science and Technology

Figure 3.17 Two shock absorbing inserts for automotive doors moulded in semi-rigid EAfoam with the ‘Foam & Film’ technology as they appear immediately after demoulding. The film protects the foaming against moisture and degradation. No release agents or mould cleaning operations are needed to produce this part in a highly automated moulding plant.

3.2.3.1 Adhesive Film Solution Basically, three types of film can be used in the adhesive Foam & Film process: • Thermoplastic polyurethane (TPU) • Polyethylene (PE) • Thermoplastic film (TP) TPU is a good film for both cold and hot processes, providing good adhesion to the foam and excellent mechanical properties but the cost is high (around 0.65 US$ per square metre). PE film can only be used in cold processes, it requires a special treatment for perfect adhesion and it gives an overall poorer performance but it has the advantage of being very inexpensive. TP film is only suitable for hot processes, it requires no treatment, and it guarantees good adhesion at a low cost (less than 0.2 US$ per square metre). This last type of film is what has been developed in depth, obtaining very good tear and impact resistance, flexibility, elongation and welding ease. This is the solution that has been used for most automated Foam & Film plants supplied so far. Using this film, General Motors makes sound-deadening parts with fully automated equipment that has a productivity of 8000 parts/day, running 24 hours per day with zero operators [17, 18].

3.2.3.2 Releasing Film Solution The second type of film, the releasing one, sticks to the mould and not to the polyurethane part. In the releasing Foam & Film technology, there are two main types of film that can be

134

Polyurethane Processing: Recent Developments used: PE and TP. PE, once again, is very common and inexpensive, but gives poor release from the foam, which minimises the benefits of this technology. TP film provides an easy process with good release from the foam for several shots at a very reasonable price.

3.2.4 Industrial Applications A good example of such a continuous moulding process that incorporates ‘Foam & Film’ technology is a system developed for the production of industrial vehicle carpets. These products are usually made out of a sandwich of two or more layers of different materials. Individually, they provide different features: aesthetics and function (a textile carpet or a synthetic mat), sound deadening (a layer of polyurethane foam) and protective (a cheap layer to protect the foam from moisture and degradation). INSOTEC, a technology designed by Cannon for the manufacture of sound-deadening automotive components, offers a variety of manufacturing alternatives that provide significant benefits such as improved quality parts, reduced costs, shorter process times, as well as regular and consistent production cycles (see Figure 3.18).

Figure 3.18 A large full-automated production plant for truck floor-covering mats using the ‘Foam & Film’ technology. The parts are made of thermoformed PVC (or thermoplastic elastomer) foam-backed with PU, which is kept separate from the mould with a film (adhesive type). 135

Advances in Urethane Science and Technology These excellent results have recently been improved with the introduction of the new ‘Foam & Film’ technology. The technology makes manual intervention unnecessary, removing the one factor, which has always been a major weakness with polyurethane processing and added a significant cost to the parts that are perceived to be an economical component of a vehicle. These components have a surface area of about 3 m2 and consist of a surface layer of polyvinyl chloride (PVC) or thermoplastic elastomer (referred to as the ‘heavy layer’), an intermediate sound-deadening layer of flexible, medium-to-low density polyurethane foam and a lower thermoplastic film. The film is designed to prevent the formation of flash during the moulding process, eliminate permanent residue usually left in the mould, and to act as a release agent (once the part is in the vehicle, this film inhibits water absorption, which is a very frequent problem with industrial vehicles). In the past, polyurethane films, which were strong mechanically and performed well but were quite expensive were used. With the new ‘Foam & Film’ technology, it is possible to replace this polyurethane film with a thinner, less costly thermoplastic one, resulting in high quality production at a lower cost. The system is composed of two shuttle-bed clamps served by one metering unit, which dispenses the pre-heated heavy layer in the mould. Pre-heating is carried out using a special infra-red heater, which incorporates easily adjustable, special low thermal-inertia resistances. When a vacuum is applied to the mould, the material adheres to the lower mould half, taking on its shape and embossed design. The protective film is automatically unrolled from an overhead source using a vertical traversing frame with pinchers that pulls an appropriate length of film over the frame and cuts it to length. The frame is positioned over the edge of the mould. Another bank of heaters slides laterally in front of the mould and warms the film to the correct forming temperature; a vacuum is applied at the end of the heating phase to conform the film to the mould. The use of two presses, as opposed to one, means that slack periods are eliminated and use of the cycle time is maximised. While the film and heavy layer are being placed in one of the moulds and pre-heated, foaming and polymerisation are taking place on the other. The polyurethane dosing and foaming section of this plant is equipped with a moulding technology that allows chlorofluorocarbon (CFC) blowing agents to be replaced with liquid carbon dioxide. This helps reduce the density of the polyurethane considerably and thus saves on material costs. The overall production time is just two and a half minutes per part. Being a twostation plant, this equipment can produce close to 50 finished parts per hour – without a dedicated operator.

136

Polyurethane Processing: Recent Developments

3.2.5 Applications This technology has already been used on turnkey plants for sound insulation parts, carpet back foaming and seat cushions. No limitations for this technology are foreseen and, actually, ‘Foam & Film’ can be implemented in almost every kind of polyurethane process needing either release agent or a cover/surface film. The system requires the presence of an open mould where the film-holding frame can be inserted to position the film prior to the vacuum-forming phase. Obviously, this technology is more easily applied to new equipment and new moulds since, most of the time, existing moulds must be modified to provide the vacuum and hold the frames. When one surface of the part is covered by an aesthetic or functional layer (carpet, plastics, etc.), obviously the film is only applied to the opposing side where the foam would be in contact with the mould (see Figure 3.19).

Figure 3.19 Turnkey installations using the ‘Foam & Film’ technology are producing numerous parts for the automotive industry.

137

Advances in Urethane Science and Technology

3.2.6 Advantages ‘Foam & Film’ offers significant benefits, such as quality parts, reduced costs, shorter process times, as well as regular and consistent production cycles. •

operator intervention is no longer required for mould cleaning and application of release agent; consequently this leads to increased productivity, uniformity and the cycle regularity which comes with a completely automatic line - a fully automated foaming process is now possible.

•

polyurethane or TP film can be used, the latter being thinner and cheaper but giving the same performance.

•

films which adhere to the product can be used, becoming an integral part of the finished component; this can be a very useful feature for non-aesthetic parts that are to be mounted in hidden positions and will benefit from this extra protection against humidity, oil, aggressive chemicals and foam-aging agents such as oxygen or other gases.

•

films, which adhere to the mould, can be re-used several times as a substitute for release agent.

•

the availability of a wide range of film sizes means no dimensional limitations on the parts to be moulded.

To summarise, ‘Foam & Film’ Technology: •

automates the production of polyurethane moulded parts,

•

eliminates mould cleaning,

•

eliminates spraying of release agent, saving its cost plus those of all the relevant dispensing equipment and special fume extraction systems (although regular fume extraction must be maintained for the polyurethane process),

•

offers a low operating cost.

3.3 InterWet - Polyurethane Co-injection The combined use of polyurethane and reinforcing agents or fillers has been common practice for a long time. A recently introduced technology - simultaneously injecting foam and glass fibre in open moulds - did not give satisfactory results, according to some of the early users. Cannon has developed an industrial solution - named InterWet - that

138

Polyurethane Processing: Recent Developments improves both manufacturing performances and mechanical properties of moulded parts [20-27]. In addition, it allows for the use of other reinforcing fibres (natural and synthetic) and opens the way to the easy addition of powder and granulate fillers.

3.3.1 Glass-Reinforced Polyurethanes, a Well-Known Technology Glass-reinforced polyurethane has been used for many years, utilizing different technologies: reinforced reaction injection moulding (RRIM), open-mould pouring over a flat mat (LD-SRIM), closed-mould injection on preformed glassmat-sandwiches (SRIM). Cannon has for many years dealt with glass-handling polyurethane technologies, developing in the early 1970s a special RRIM mixing head able to mix formulations containing high percentages of milled fibre, and launching in the 1980s the HE metering machines, closed-loop electronically-controlled piston-metering systems able to cope with the abrasion deriving from glass and mineral charges. Later in the 1980s the Compotec preformers were introduced, for the production of preformed glass mats required by SRIM - structural-moulding applications. A new technology has been introduced by Krauss-Maffei, long fibre injection (LFI) [28], that simultaneously pours polyurethane and chopped glass roving on the surface of open moulds; after the shot the press is immediately closed to allow for the expansion of the foam, which surrounds all the fibres and produces a lightweight, resistant composite panel. Most recently, Hennecke has introduced FipurTec [29, 30], a technology where the fibre is chopped outside the head and projected into the flow of reacting chemicals, where this touches the surface of the mould. In essence these systems carry out a job similar to that performed for very many years in the polyester ‘chop-and-spray’ applications, the only differences being that polyurethane has a different profile of reactivity and viscosity build-up than polyester and that, at least in one case, the glass roving is conducted through the mixing device instead of being fed outside of it. The interest behind this technology lies in the fact that a thin, resistant composite part can be moulded in only one operation. Using preformers requires more equipment, space, investments, etc. In addition, glass roving costs roughly half that of glass mat, and this can represent a considerable saving when producing large, highly reinforced parts.

3.3.1.1 Practical Problems That Needed to be Addressed A few practical problems have been identified in this technology, mainly deriving from the design of the mixing equipment currently available. In one case the chopped fibre is

139

Advances in Urethane Science and Technology fed through a pipe that is co-axially positioned in a conventional ‘L-shaped’ head and then the mixed resin is conveyed around the pipe that feeds the chopped fibre. In the other case the fibre is projected separately from the polyurethane; in this case the glass meets the polyurethane only at the end of the head’s final discharge duct. The separation of the fibre and polyurethane before the exit of the head results in wetting of the fibres not being optimal; this is visible during moulding, where part of the glass gets stuck vertically into the base of rising foam, visibly dry. The main problem comes from the dry chopped glass that - leaving the nose of the head still not wet - flies everywhere near the mould surface. This results in negative effects on workers, on cleaning and maintenance operations, and on part quality. This problem can seriously hinder the application of this technology.

3.3.1.2 The Cannon Approach Cannon started looking into this technology in the summer of 1997, at the urging of some car parts producers who were unsatisfied with the performances of the existing ones. The main objectives of the project were: •

To supply in the shortest possible time a reliable solution for the injection of glassreinforced polyurethane foams: with as little as possible an impact on the working environment, with the best possible quality,

•

To provide the lightest possible mixing/dispensing equipment, in order to reduce the investment in head-carrying robots,

•

To design a multi-purpose solution, not limiting the choice of charge to glass roving only, but including a wider range of natural and artificial fibres, as well as pulverised fillers,

•

To provide maximum flexibility in charge feed, so that products with different content of charges - in the same moulding or between subsequent moulds - can be produced,

•

To provide maximum number of automatic checks on the charge’s feeding line, to avoid the troubles experienced during automatic production cycles.

Due to the short delivery time required, it was not realistic to conceive totally new equipment for this application, and it was decided to concentrate on the use of existing, proven pieces of hardware. Performance of the foaming section had been optimised in previous years, and the main obstacle was the proper handling of glass and other fillers.

140

Polyurethane Processing: Recent Developments

3.3.1.3 The Technologies Involved The project was split into three parallel lines of development, to speed up results and make use of the experience of various existing specialists for the specific engineering tasks: •

Charge/Reinforcement (handling, dosing and feeding)

•

Polyurethane (dosing, mixing and laydown)

•

Robotics.

Handling Reinforcements and Charges This project required a multi-functional solution, able to accommodate various types of fibres, fillers and charges; accordingly, the design of the whole handling section would have had to respect this future necessity. Since the most urgent need called for the use of glass roving, the first development focused on the design of a simple, reliable device to store, meter and chop conventional type glass-roving rolls. A proper technical specification was drafted to provide optimum handling of the types of glass most suitable for this peculiar application. Various qualities of roving are available on the market for different end-uses: the diameter of the basic threads (bunches of individual glass fibres) and of the final roving (the rope made with various threads) define the weight and the field of application, while the type of coating (a thin layer of special resin applied on the fibres after the glass-extrusion phase) determines the best compatibility of the roving with the polymeric matrix. A specific type of Owens Corning Fibreglass (OCF) roving was selected as the reference, having a 2400 Tex specific weight, i.e., the roving weighs 2,400 grams per kilometre, with high ‘softness’ of the thread. Other types would have been easier to cut (or to be broken, since a fibre of glass is not shear-cut - it is bent beyond its critical radius until it breaks), but their individual fibres were more agglomerated in the thread and would have been more difficult to wet. All parts were designed to cope with a wider range of roving. The glass-feeding device was conceived to be installed over the polyurethane mixing head, and its design was made in strict cooperation with the team in charge of designing the head. It is basically composed of: •

A pulling/cutting device: two opposed rolls that rotate and trap between them one or more roving, pulling them from their storage. A blade - held in one of the two rolls forces the fibres to bend beyond their critic radius until they break;

•

A conducting system: a flow of compressed air that pushes the chopped fibre from the cutting place to the mixing head.

141

Advances in Urethane Science and Technology A few interesting technical features characterise this roving feeding-chopping section: •

A hydraulic motor, to drive the glass-roving pulling rolls, was chosen to provide proper speed, torque and easy variation of the driving speed. Its high power/weight ratio made it preferable to an electric motor;

•

A special cutting device was conceived, based on a rotating blade-holder (very easy to access and maintain) capable of accommodating up to four blades. By fixing more than one blade in the proper holders one can chop the roving in different lengths, the longer size being that equal to the circumference of the rotating blade-holder. The part had to be dimensioned to resist the very high operating speed, up to 5,000 rpm, without suffering from the presence of glass debris around rotating shafts. This design proved useful later, in the further development of a variable-length chopping system;

•

A glass-guide system using compressed air was designed to take away all the cut roving from the chopping device and bring it through the mixing head down onto the mould, with the additional task of providing a final purge of the discharge duct.

The Mixing Head Most of the development work was concentrated in designing a mixing head able to wet all the solid components before they left, to ensure the projection of a very well wetted mixture rather than a blend of liquid only coupled with a dry, flying filler. The mechanism which could ensure a thorough wetting of high percentages of solid fillers with a mixture of polyurethane is to provoke high turbulence in the mixing head’s area where liquid and solid meet, and then reduce the turbulence to allow for splash-free open-mould distribution of the blend obtained. It was in fact observed that, with heads where the junction between liquid and solid was in a turbulence-free zone, the wetting was not optimal. By using the Cannon FPL head, an existing, two-decades-proven, piece of hardware it would have been possible to obtain the best available mixture of the polyurethane component with a turbulence-free laydown in the mould. Its internal geometry is L-shaped, with two cylindrical chambers of different diameters connected at a 90º angle. The turbulent flow created by impinging of two components in the small mixing chamber is quickly converted into a smooth laminar flow as the chemicals are diverted through 90 degrees downwards into a large discharge duct and then leave the head. At the end of the pour the smaller mixing chamber is sealed and automatically cleaned using a hydraulic piston. By using the original mixing chamber for the liquid components and boring a hole through the main cleaning piston - the plunger - so that a solid component could be fed through it, a wide range of fillers can be used (see Figure 3.20). When the mixing operation begins, the 142

Polyurethane Processing: Recent Developments

Figure 3.20 The Cannon FPL head used for InterWet incorporates in its upper section the glass cutting mechanism and relevant safety checks.

main piston is retracted until it clears the discharge duct at the junction of the two cylindrical chambers. In a controlled sequence, the small piston sealing the mixing chamber retracts and the two chemical components are fired at high pressure against each other through injectors in the mixing chamber before leaving it through the discharge duct. The metering and feeding operation of the solid component from the head’s upper part is controlled mechanically, assisted by an intense flow of compressed air. Its feed is synchronised with the arrival of the liquid blend from the mixing chamber. The solid component meets the liquid blend and together they are co-injected into the mould. Following co-injection, the small piston seals the mixing chamber, the large piston seals and cleans the discharge duct and the feed of solid component is interrupted, although the flow of air is maintained for a short time in order to clean any residual polyurethane - which could interfere with the next pouring operation - from the nose of the head.

143

Advances in Urethane Science and Technology

Figure 3.21 Open-mould co-injection of PU and glass fibre is executed at high speed and in a clean working environment, with an InterWet machine.

The innovative concept in this solution is that the solid component meets the liquid formulation just in front of the mixing chamber, where the turbulent flow of the mixture is directed through a 90° angle downwards creating a laminar flow. In this way, the kinetic energy from the pressurised liquids is used to wet the stream of solid component thoroughly and efficiently at a point just 20-25 mm from the point of impingement. The blend of polyurethane and filler leaves the head already well blended, and there is no evidence of ‘flying glass’ out of the head (see Figure 3.21). This special design and technique ensures that the solid component is thoroughly internally wetted, hence its name, InterWet.

144

Polyurethane Processing: Recent Developments Robotics The development of a proper device to carry the mixing head during the pouring operation over open moulds required relatively short time. Basic requirements were multi-axes capacity, speed, precision and easy programming. The designers concentrated on the use of commercial polar robots, operating up to 6 axes at high speed and precision (see Figure 3.22). The limited weight of the Cannon FPL mixing head allowed for the selection of a mediumsized model able to carry on the wrist 125 kg of payload at up to 2 m/s, with acceleration of 3 m/s2. This robot was fitted with two storage boxes for the roving, mounted on the elbow area (actually this model allows to install there up to 3 glass-roving rolls, for a maximum 60 kg payload). This is a very convenient glass-storage solution for laboratory and small production applications. When high quantities of glass are required, a solution able to reduce rolls-replacement times must be provided. A proper device has been designed to accommodate the 1,000 kg of glass demanded for industrial heavy-duty tasks. In this case the glass rolls - that could be cascade-joined one another - must be positioned the closest possible to the head, so that the way the roving must take to reach the chopper is not restrained by long guides.

Figure 3.22 One of Cannon’s R&D units devoted to the development of InterWet technology.

145

Advances in Urethane Science and Technology First Results The assembly of the pieces of hardware was completed in less than eight weeks from the beginning of the project, and immediately trials were scheduled in the laboratory using a flat test-plate mould. A conventional Cannon HE - a piston-driven, electronically closed-loop controlled dosing unit available in the laboratory - was connected to a new mixing head, that had been fitted to the wrist of the robot by means of proper flanges. The glass was carried from its storage to the chopper by means of flexible plastic pipes, and the path was designed so that there would be no obstructions and friction for the roving. Glass roving X900A, a 2400 Tex, very soft roving with average attitude to the cut, was supplied by Owens Corning Fibreglass. A chemical formulation supplied by Dow, characterised by cream time of 12 seconds, gel time of 40 seconds, demoulding time of 3 minutes and free rise density of 50-55 kg/m3 was used. The first tests were run using approximately 30 parts of glass over 100 parts of polyurethane. Surprisingly, the results were extremely satisfactory from the very beginning (see Figure 3.23).

Figure 2.23 Homogenous wetting of all fibres was obtained since the very first trials with InterWet, thanks to the interval geometry of the mixing head. 146

Polyurethane Processing: Recent Developments After fixing minor problems with the cutting control, an optimum distribution of blend was obtained, and the glass fibres were all wet (this was clearly visible because polyurethane was pigmented in black, and absolutely no white spot of dry glass was shown at the end of an open-mould pouring operation). Being all wet, the fibres did not tend to stick vertically in the rising foam, therefore less possibility remained for the air to remain trapped in the mixture: the moulded parts were, in fact, totally free of air bubbles. A number of trials were satisfactorily run using thin, transparent polyethylene film as liner for both mould halves, so that when the liner was peeled off it was possible to evaluate the quality of the compound and its surface (see Figure 3.24). The second series of trials was run on a mould for production of automotive door panels where a layer of expanded or compact PVC sheet was unrolled on the lower mould-half and cut to measure, its edges were blocked with a frame and it was vacuum-formed to perfectly adhere to the mould surface. After this operation the blend of polyurethane and glass was poured on the rear side of the PVC liner, and the mould carrier closed for polymerisation. On this occasion - since the first results were very encouraging - the

Figure 3.24 No holes are visible in this InterWet non-presses laboratory sample, photographed against a strong light. The diffuser-mounted on the InterWet provides a very distribution of material all over the mould.

147

Advances in Urethane Science and Technology percentage of glass was increased to 35% of the total blend, i.e., approximately 60 parts of glass to 100 parts of polyurethane. Again, very nice parts were obtained at the early stages (see Figure 3.25).

Figure 3.25 Nice parts, free from air entrapments between reinforced PU and outer skin were obtained at a very early stage of the InterWet development work.

First Improvements •

Improved distribution [25]

One aspect of the laydown - the width of the pouring path - was worth immediate attention; since the reactive blend leaving the head covered only a few centimetres of mould at each pass, a rather high number of passes was required to properly cover the whole surface which forced use of the robot at very high speed to complete the whole pattern before creaming started. To avoid the overlapping of two contiguous layers of mixture, sometimes a narrow gap remained between the layers. This did not create problems but the distribution of glass in those thin strips was probably different than in the other areas of the mould.

148

Polyurethane Processing: Recent Developments To overcome these problems, a specially-designed pneumatic device was added at the exit of the mixing head, that deviates, under controlled conditions, the flow of mixture from its natural, vertical direction and distributes it over a wider pattern and, consequently, in a thinner section. By doing so it is irrelevant if two contiguous layers do overlap (since the layers are thinner the effect is negligible) and fewer passes are required to cover the same surface. The robot can be used at lower speed (or a simpler, less expensive robot can be used) or more parts can be made in the same unit of time. This could be relevant thinking to carousel-based operations, where the robot would have to perform different pouring patterns in a sequence, on different moulds. Another advantage derived from this pneumatic distributor is that the jet of air pushes even further down the fibres, which results in a very thin layer of mixture applied on the mould surface. This means that the male half of the mould - when closing over the female part - does not ‘wash down’ the vertical walls from the layer of reacting foam and glass which covers them, since this layer is quite thin and does not constitute an obstacle for the lowering plug. Parts as thin as 1.5 mm can be moulded with optimum surface aspect and homogeneous distribution of glass across the whole surface.

•

Variable fibre length, with ‘on-the-fly’ change

The first fibre-cutting device, as stated above, had the possibility of accommodating more blades. Therefore, the length of cut fibre obtained was equal to the circumference of the cylindrical blade-holder when only one blade was installed, or half of its circumference with two blades, or one-quarter of it with four blades. Since this was foreseen in the original wish list, the option of allowing the user to vary the length of glass fibre at will, even ‘on-the-fly’ on the same moulding, was soon taken into consideration. A simple, sturdy mechanical device was designed to allow for this variable cutting. This mechanism which operates via an impulsion given by the robot’s programme at a given point of the pouring pattern, cuts the roving in shorter sections; this variablelength cutter provides different cuts of fibre: from L to L/12, where L is the longest possible choice and the combinations depend on the use of the 12 blades installed on the holder.

•

Variable glass fibre output

Operated by a hydraulic motor, the pulling device (two opposed rolls that rotate and trap between them one or more roving, pulling them from their storage) is able to provide variable output of glass. By electronically controlling an inverter or a proportional valve it is quickly possible to vary the motor speed, providing more or less fibre in different

149

Advances in Urethane Science and Technology parts of the moulding. The motor can be even stopped (thus feeding only polyurethane foam) or left running after the end of the polyurethane shot (providing dry, unblended glass fibre over some areas of the mould where an extra reinforcement would be desired).

•

Extended safeties

The main concerns in operating this kind of plant derive mostly from the irregular feed of glass to the head. For this reason the latest InterWet machines have been equipped with detectors able to command an immediate stop to the machine in case of: •

Empty storage of glass roving

•

Roving blocked in its roll

•

Roving blocked somewhere along the feeding line

•

Roving not properly fed into the head after the cutter and forming a ‘birds nest’

•

Malfunction of a single blade.

3.3.1.4 Advantages of the InterWet Process These new developments have resulted in a higher degree of flexibility for the InterWet machines whose advantages versus the existing and available similar methods can be summarised as follows: •

Optimum overall mechanical properties, due to an optimum glass fibre wetting with PU, because of the mix-head design

•

Better distribution of the mechanical properties of the moulded part, i.e., no areas without or with little glass, due to a more homogeneous glass fibre distribution in the mould, because of the special distribution system.

•

Programmable variation of mechanical properties on the same part, since the distribution of the composite’s mixture on the surface of the mould is fully programmable. One can have in different areas of the same part:

•

PU and glass in fixed proportion

•

PU and a varying proportion of glass

•

PU only

•

Extra glass over an already filled area.

150

Polyurethane Processing: Recent Developments •

Easy operation because it uses a standard (modified) FPL mix-head with obvious consequences in terms of simplicity of maintenance, reliability in use and availability of spare parts when required.

•

Flexibility in operation because it is designed for more than one type of additive component: it can easily be converted to process fillers or fibres or chopped scrap foam. Its range of application includes continuous reinforcing fibres, natural fillers (mineral or organic), pulverised plastics, etc.

•

All the operational advantages deriving from the use of a light mixing head (robot size, speed, productivity, etc).

•

Economics, because - being based on commercially available components - its mechanical parts can be less expensive than specially developed ones.

When compared with other glass-reinforcing technologies the following advantages must be added: Versus some RRIM applications: •

Economics, since it is no longer necessary to carry out the expensive dispersion of solids in the polyol component - now, by adding them directly in the mixing head, it is possible to eliminate all the traditionally associated problems, i.e., possible absorption of polyol into the filler with the resulting slow release of reactive components in the polymer matrix, abrasion, clogging of lines and tanks, fluctuating percentages of filler in the liquid, etc.

Versus some mat applications: •

Economics: in glass (roving costs less than mat) in equipment (one more machine, storage, handling, floor space, etc.) in scrap (preformed parts - although optimised - require punch-cutting that produces scrap trim) in stock (stored preforms can be damaged by sunlight, dust, moisture, and often must be thrown away) in manpower (to handle, position, punch, etc.)

151

Advances in Urethane Science and Technology •

Quality (foam mixes better when co-injected with glass roving rather than when injected into a pressed mat preform)

•

Flexibility: it’s easy to mount on a carousel various moulds and produce small series exploiting the flexible properties of the robot.

It is evident that there are applications where both RRIM and mat preforms are still preferred because of their specific advantages. InterWet has been conceived to fill the gap between the two technologies, and allow for more extended use of reinforcements and fillers in applications where neither RRIM nor mat preforms could be used [31].

3.3.1.5 Applications When this technology was first introduced it immediately appealed - for the above mentioned reasons – to automotive and transportation end users. Volumes, cycle times and final properties meet this market segment’s requirements, but until the currently available solutions address some of the above-illustrated problems this industry would not select it for mass production. A market segment where this application would be really interesting is the manufacture of special sanitary-ware parts, for at least three different purposes. The quickest to understand for its immediate advantages is the combination of an outer plastic shell with either aesthetic or functional purposes - back-foamed with reinforced polyurethane; examples include whirlpool tubs, bath tubs, shower trays, and vertical elements for shower cabins integrated with accessories. These products are characterised by medium-small production series, wide range of models, and large surfaces. The high operating flexibility allowed by the programmable robot is very adaptable for the production of this type of moulding: the larger the part, the more advantages can be found in this technology. Medium-deep-draw shaped parts can be easily thermoformed from large thermoplastic sheets (even in medium-small series) and immediately reinforced in an integrated production line including thermoformer, foaming station and curing area. Dimensions, productivity and investment must be tailored to the project’s requirements. Another area of interest for this new technology is the re-use of chopped foam scrap, either flexible or rigid; the design of the solid component’s feeding device - as foreseen in the project’s objectives - has been made in a way that, using an appropriate volumetricfeeding mechanism, powders, granulates and recycled foams can be dosed in the mixing chamber through the same channel bored in the mixing head’s cleaning plunger. It is possible to add roughly pulverised scrap foam to the virgin formulation directly in the mixing head. This would solve a recycling problem, as well as lowering the cost of

152

Polyurethane Processing: Recent Developments

Figure 3.26 In addition to glass roving a number of other materials can be co-injected with PU using the InterWet technology, scrap PU foams, sand, ground cork and other fillers that would be difficult to add in the component tank.

formulation. Positive trials have been made in a development phase, with filler loading as high as 10% of the final part’s weight (see Figure 3.26). Another interesting potential application is the use of continuous fibres, cut in mediumlong size, to reinforce thin pads and cushions moulded with flexible foams. The continuous development of lighter, thinner parts - where weight-reduction and volume-exploitation is the object of intense development - requires the use of formulations characterised by very high mechanical properties that unreinforced foam cannot meet. The use of long natural or synthetic fibres to reinforce these flexible mouldings had applied for a long time, but only under the form of textile inserts manually inserted in the mould prior to the foaming operation. Being able to co-inject foam and long fibres - applying longer or shorter sizes according to the mechanical properties required - represents an advantage that comes at a very attractive investment cost.

Note: The Cannon FPL Mixing Head is covered by several international patents.

153

Advances in Urethane Science and Technology

References 1

M. Taverna and B. Pile, Presented at the Polyurethanes Expo ’99 Conference, Orlando, FL, USA, 1999, p.37.

2

Modern Plastics International, 1999, 29, 3, 127.

3

M. Taverna, Presented at the Recycle ’94 Conference, Davos, Switzerland, 1994, Paper No.35.

4

L. White, Urethanes Technology, 1995, 12, 2, 3.

5

Plastics and Rubber Weekly, 1995, No.1612, 10.

6

D. Smock, Plastics World, 1995, 53, 11, 20.

7

British Plastics and Rubber, 1995, November, p.10.

8

D. Smock, Plastics World, 1995, 54, 10, 32.

9

M. Taverna, Presented at the Utech Asia ’97 Conference, Suntec City, 1997, Paper No.12.

10

D. Smock, Plastics World, 1996, 54, 9, 23.

11

C. Fiorentini, M. Taverna and J. Luca, Presented at the Polyurethanes ’95 Conference, Chicago, IL, USA, 1995, p.476.

12

Plastiques Flash, 1998, 304/5, 92.

13

M. Taverna, P. Corradi and B. Biondich, Presented at the Polyurethanes ’95, 1995, Chicago, IL, p.484.

14

M. Taverna and P. Corradi, Presented at the Utech Asia ’96 Conference, The Hague, 1996, Paper No.10.

15

P. Mapleston, Modern Plastics International, 1997, 27, 8, 30.

16

Plastics and Rubber Asia, 1997, 12, 72, 34.

17

British Plastics and Rubber, 1997, July/August, 19.

18

Plastics and Rubber Weekly, 1997, No.1687, 17.

154

Polyurethane Processing: Recent Developments 19

M. Taverna, Presented at the Polyurethanes Expo ’98 Conference, Dallas, TX, USA, 1998, p.685.

20

P. Mapleston, Modern Plastics International, 1997, 27, 11, 41.

21

Plastics Rubber Weekly, 1998, No.1729, 13.

22

Macplas International, 1998, February, 46.

23

G. Graff, Modern Plastics International, 1998, 28, 11, 38.

24

British Plastics and Rubber, 1998, November, 20.

25

M. Taverna and A. Bonansea, Presented at the Polyurethanes Expo ’98, 1998, Dallas, TX, 687.

26

European Plastics News, 1999, 26, 4, 48.

27

Composites – French/English, 1999, 36, 50.

28

J. Stark and F. Peters, Presented at the Utech ’99 Conference, 1999, Singapore, Paper No.6.

29

Plastics and Rubber Asia, 1997, 12, 70, 34.

30

D. Smock, Plastics World, 1996, 54, 12, 35.

155

Advances in Urethane Science and Technology

156

4

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels for Super Insulation Applications Paolo Manini

4.1 Introduction Rigid polyurethane (PU) foam is the preferred insulator material in a wide range of applications encompassing household, industrial and commercial appliances (refrigerators, freezers, display cases, vending machines), transportation (refrigerated trucks and reefers, shipping containers), insulation in buildings and in industrial plants. The replacement of chlorofluorocarbons (CFC) with environmentally more benign chemicals, as recently mandated by the Montreal Protocol and subsequent revisions, has caused a reduction of the insulation efficiency of the PU foams, since the new blowing agents available, like hydrocarbons, hydro-fluorocarbons (HFC) or carbon dioxide have worse gas thermal conductivity properties. In applications, the lower insulation efficiency of the CFC-free foams generally leads to higher energy consumption levels unless specific compensation actions are taken. This issue is particularly severe for the household appliance industry, which has also been called upon over the past years to progressively reduce the energy consumption of their products. With the energy efficiency targets getting more and more demanding and only marginal improvement obtainable from the optimisation of conventional technologies, the appliance manufacturers are seriously looking at alternatives which may provide additional benefit in terms of energy savings. The appliance industry is not the only one involved in the debate on energy consumption reduction. The growing attention to the global warming issue is forcing policy makers to take actions to cut carbon dioxide emissions and minimise the greenhouse effect. With this in mind, the Kyoto Conference has posed ambitious targets for energy saving, the achievement of which calls for improvements in many fields, from transportation, to refrigeration and building insulation. The ban on CFC and the demand for a more efficient use of energy, both issues coming from environmental concerns, are therefore posing serious challenges to the Industry as a whole.

157

Advances in Urethane Science and Technology Vacuum Insulated Panels (VIP), having a thermal conductivity three to five times lower than conventional PU foams, allow the achievement of superior insulation performance. They can be used to partially replace the conventional insulation materials to provide a more efficient insulation structure, which allows energy saving without the need to increase the insulation thickness. Alternatively, VIP can be used in those applications where it is important to reduce the insulation thickness to a minimum value without loosing thermal performances. A VIP is obtained by packaging a microporous low conductivity filler material inside a highly impermeable gas barrier bag. The filler is then evacuated to a proper vacuum level and the bag sealed. A gas adsorbent, normally referred to as a getter, inserted in the bag before sealing, is also necessary in most panel designs to ensure the proper vacuum level during the lifetime of the panel. Several fillers have been proposed in the past, either in the form of compressed powders or fibres, however all have some disadvantages in terms of cost, process complexity and/ or weight. This has prevented VIP technology from finding widespread acceptance in most applications. The recent development of open cell rigid foams, first introduced into the market in the early 1990s has sparked off new interest in this technology. However, in order to fully exploit the unique properties of these insulating materials, the reevaluation and re-design of the other key elements of a vacuum panel, i.e., the barrier film, the gas adsorbent and the manufacturing cycles, has also been necessary. This process, which is still ongoing and has potential for further improvements, has required the commitment of several sectors of the industry in the last five to six years. In this chapter, the present status of the open cell PU foam-filled vacuum panel technology is discussed and some recent developments in film, getter and processing technologies are reviewed. Special focus is given to the vacuum issues, which are key for the proper selection and treatment of the VIP components. Some specific aspects, related to the manufacturing, characterisation and practical use of vacuum panels, as well as the achievable benefits, are also considered. Examples of applications where this technology is finding its place in the market and areas where further work is necessary to make vacuum panels more cost effective are presented.

4.2 Some General Properties of Open Cell PU Foams for Vacuum Insulated Panels Vacuum panels have been studied as a means to improve thermal insulation for a long time. Several insulating fillers, such as silica and perlite powders [1, 2, 3], fibre glass [4, 5], and aerogels [6, 7, 8] have been proposed as core materials for VIP, each of them

158

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… having specific advantages and disadvantages in term of weight, thermal conductivity, handling, processing, vacuum properties and cost. To overcome some of the drawbacks associated with the use of these materials, new families of fillers, based on PU [9, 10, 11, 12] and, more recently, polystyrene (PS) [13] open cell foams, have been developed and proposed to the market. Both open cell PU and polystyrene foams are now being considered as interesting options due to their moderate outgassing, good thermal insulation values, low weight, ease of handling and cost effectiveness. This chapter will focus on the open cell PU foams, even though many general aspects and recent developments of the technology, dealing with films, getters and manufacturing processes, can be also applied, with minor changes, to other micro porous fillers. The thermal conductivity, or λ factor (mW/m-k), for some of the most popular filler materials is shown in Figure 4.1 as a function of pressure (Pa). The thermal conductivity of cyclopentane (CP) blown closed cell PU foam is also given for comparison.

Figure 4.1 Thermal conductivity as a function of pressure for some filler materials

159

Advances in Urethane Science and Technology Regardless of the type of filler, all curves have a similar behaviour, with a low pressure region, where the λ factor is constant, followed by a region where the λ factor increases with increasing pressure. This common trend can be explained considering that the total apparent thermal conductivity in micro porous materials is the sum of four physical contributions, i.e., the thermal conductivity through the solid, λs , the thermal conductivity through the gaseous phase, λg, the thermal conductivity by thermal radiation, λr and by gas convection, λc: λ = λs + λr + λg + λc

(1)

The last term of Equation (1) can be neglected for core material with cell size smaller than 1 mm. An extensive treatment of the above physical mechanisms and their role in heat transfer in low density closed cell foams has been provided by Glicksman [14]. In the case of the open cell PU foams detailed expressions for the first two terms of Equation (1) have been given by Kodama and co-workers [9]. Both terms are independent of pressure and strictly related to the foam density and morphology (cell size, structure and degree of anisotropy). In particular, λs is directly proportional to the thermal conductivity Ks of the foam and inversely proportional to the cellular anisotropy η according to the formula [9]:

[(

) ]

λ s = 1 − Vg2 / 3 K s / η

(2)

where Vg is the fraction of foam volume occupied by the gaseous phase. To reduce the heat transfer through the solid it is therefore necessary to increase the cellular anisotropy and to lower the thermal conductivity of the polymer structure. Following Kodama and co-workers, the heat transfer by radiation is given by: λ r = 2 / 3 Vg1/ 3 H r d

(3)

where d is the cell size and Hr is the coefficient of thermal conduction by radiation which depends on the emissivity of the polymer and the morphology of the cell. To minimise the λr contribution, the cell size and Hr have to be reduced. The former can be achieved by adjusting the formulation of the foam, the latter by reducing the opening size of cell membranes to make radiation transmission through the cellular windows less efficient.

160

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… The heat transfer by gas conduction, λg, becomes progressively more important as the pressure increases, as shown by the following set of equations [15]:

(

)

λ g = λ goΠ / (1 + 2L / d)

(4)

where λgo is the thermal conductivity of air at atmospheric pressure, Π is the porosity of the foam and L is the mean free path of air, linked with pressure P by: L = (kT ) / ⎛ 2 π P σ 2 ⎞ ⎝ ⎠

(5)

where T is the absolute temperature, k the Boltzmann constant and σ the collision diameter of air (about 4 x 10-10 m2). The λg contribution vanishes for L>>d. For currently available foams, having an average cell size in the 100-200 μm range, this means that pressure has to be kept at 1.0 Pa or below to completely eliminate the heat transfer by gas conduction. To ensure this condition the foam has to be properly evacuated and all gas sources deteriorating the vacuum during the panel life effectively compensated by a suitable adsorbent, or getter, having sufficient gas capacity and efficiency. As shown by Equations (1-5), the insulation properties of an open cell foam depend on a complicated interplay of different parameters such as the PU thermal conductivity, foam density, cell size and cell size distribution, cell morphology and anisotropy, which have to be optimised during the foam preparation to obtain the best trade-off among the three thermal conductivity contributions, λs, λr and λg. In this process, the mechanical properties of the foam cannot be neglected, being essential to ensure structural stability of the vacuum panel. In fact, after sealing, the vacuum panel is exposed to the hydrostatic load of the atmospheric pressure and has to withstand it for a long time, which can be 15-20 years or even more depending on the application. Dimensional stability tests carried out on open cell PU foam-filled vacuum panels showed that creep problems should not occur for properly prepared panels, provided the open cell foam preparation has been optimised [15]. The open cell PU cores are produced according to various preparation technologies, such as lamination and block foaming. In both cases, after the foam has been grown, a cutting operation is usually necessary to remove the outer skin of the material, which

161

Advances in Urethane Science and Technology contains a high percentage of closed cells, and to slice the PU into slabs of the proper size and thickness. The removal of the outer skin may be a labour intensive process which produces waste, lowers the process yield and increases the overall production costs, this issue being particularly important in the case of the lamination process. Great attention is therefore now being paid to design and the open cell foam production process, to reduce waste and increase productivity. A different approach to making open cell foam slabs has been proposed recently [11], which is based on the use of the PU fluff obtained from the recycling of used refrigerators (Recycled Urethane Fluff, or RUF panel). During the recycling process, the rigid insulation foam contained in the old refrigerators or freezers is mechanically ground into a fine powder, to completely separate and recover the CFC present in the foam. The resulting fluff, composed of completely open cells, is dried and introduced into a binding machine where it is sprayed with a certain amount of isocyanate (15-20% by weight) and thoroughly mixed. The blend is then transferred to a mould, heated at 120 °C and compressed at 0.5 MPa for 10 minutes to consolidate it and remove residues of the process. After de-moulding and curing, the RUF panel can be used as core material in a VIP. The fluff obtained from recycling one single refrigerator can be used to produce enough vacuum panels to insulate a new appliance, thus generating a virtual recycling loop. Due to the use of very fine compressed powder, the density of this open cell foam is three to four times higher than those produced by lamination or block foaming. Selected physical properties of some open cell PU foams are listed in Table 4.1.

162

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Table 4.1 Some physical properties of open cell PU foams Ref. [56]

Ref. [56]

Ref. [9]

Ref. [12]

Ref. [11]

Lamination

Block foam

-

Block foam

RUF

60-64

52-58

52

60-65

180-200

100

130-170

140

-

-

Perpendicular

-

-

75

-

-

Parallel

-

-

205

-

-

Thermal conductivity (mW/m-K) @ 5 Pa

7

5.5-7.5

4.5-5

7

8-10

Production process Density (kg/m3) Cell size(μm) Average

Compressive strength (MPa ) Average

0.25

0.23-0.35

0.26

-

-

Parallel

-

-

0.383

-

-

Perpendicular

-

-

0.133

-

-

Pre-treatment

120 °C x 30 min

120 °C x 30 min

140 °C x 1 hour

120 °C x 30 min

120 °C x 30 min

100

100

10 0

100

100

Open cell content (%)

4.3 Vacuum Issues in the Selection of VIP Components 4.3.1 Vacuum Properties of the Open Cell Foams To fully exploit the insulating performances of the open cell PU foams, pressure in the panel has to be kept preferably below 1.0 Pa during its life. To achieve this demanding target, the foam must be 100% open celled and with a very low outgassing rate. In spite of the increasing use of open cell PU foam, data from the literature on its vacuum properties are quite scarce. To estimate the outgassing contribution, specific tests can be carried out using high vacuum benches equipped with a quadrupole mass spectrometer.

163

Advances in Urethane Science and Technology

Figure 4.2 Scheme of the stainless steel high vacuum bakeable bench equipped with mass spectrometer for outgassing tests on vacuum components. IG: ionisation gauge.

To quantify the relatively small amounts of gas species released from the sample, it is necessary to use very clean, bakeable, high or ultra-high vacuum systems made with stainless steel and/or glass components. This ensures better leak-tightness and a negligible gas emission from the bench. A sketch of a typical experimental apparatus, having two separate pumping groups and base pressure at 1 x 10-6 Pa or lower, is shown in Figure 4.2. A practical technological implementation of the scheme shown in Figure 4.2 is illustrated in Figure 4.3. The sample to be analysed is mounted in a glass bulb which is connected to the test bench and evacuated for 10 minutes with a turbo and a rotary pump to a final pressure of 10-4 Pa. The bulb is then isolated from the pumps and the total pressure increase, due to the sample outgassing, is monitored with a capacitance manometer for some days. The use of the capacitance manometer avoids changes in the gas composition which might occur if a hot filament pressure gauge is used. From time to time a small amount of gas is sampled from the test volume, through valves V4 and V6, and passed to the mass spectrometer for partial pressure measurement. Before running the outgassing tests

164

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.3 A stainless steel outgassing bench equipped with a quadrupole mass spectrometer and an external oven for vacuum bake-out (Reproduced by courtesy of SAES Getters SpA, Italy).

on the PU sample, a blank run has to be carried out. The result of the blank run is then subtracted from the outgassing test to get the actual gas emission from the PU sample. The use of two distinct pumping groups is essential to minimise the system contamination (mainly water vapour) which occurs when the bulb is opened to the air, to mount the specimen. In fact, during this operation, valve, V4, is kept closed, reducing the surface area of the bench which can absorb water and atmospheric gases. Before being used in a vacuum panel, the open cell PU foam needs a preliminary heat treatment in air, generally carried out at 120-150 °C for 10-60 minutes to remove water and other volatile species which otherwise would desorb and rapidly cause the vacuum to deteriorate. The result of a typical outgassing test carried out at 23 °C on a foam sample baked at 120 °C for 30 minutes is shown in Figure 4.4 for all desorbed gases but water. Water is difficult to quantify since it sticks to the walls of the system and only partially reaches the mass spectrometer. Water can be estimated as the difference between the total absolute pressure and the sum of the partial pressures of the other gas species, which can be accurately quantified with the mass spectrometer.

165

Advances in Urethane Science and Technology

Figure 4.4 Outgassing rate for an open cell PU foam. Water is not shown.

For each gas species, the amount released by the sample at a given time is obtained by simply multiplying the test volume (V) by the gas partial pressure (qi(t) = Pi V, qi(t) is the outgassing rate (Pa-l/h) at time t, Pi being the measured i-th partial pressure). The experimentally determined qi value for each gas species can be interpolated over time according to the semi-empirical law quoted in the literature [16]:

q i (t) = q0i (t − αi )

(6)

where qi(t) and q0i are the outgassing rate (Pa-l/h) at time t (hour) and the desorbed amount (Pa-l) after 1 hour of the i-th gas species, respectively. The dimensionless parameter αI is related to the desorption mechanism of the i-th gas species, its value usually ranging from 0.5 to 1, depending on the gas species and the material considered [16, 17]. Integration over time of Equation (6), provides an estimate of the PU sample outgassing load Qi for each gas species after a given time. As an example, the estimated gas released after 1 to 20 years for the sample of Figure 4.4 and for a 50 x 50 x 2 cm3 panel size is given in Table 4.2. The main gases, besides water, are carbon dioxide, nitrogen, carbon monoxide and hydrogen. It is interesting to note that, given the nature of the desorption

166

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Table 4.2 Outgassing parameters and estimated gas load Qi (Pa-l) after 1, 5 and 20 years for a 50 x 50 x 2 cm VIP Gas

q0i (Pa-l/h)

αi

Qi (1) (Pa-l)

Qi (5) (Pa-l)

Qi (20) (Pa-l)

CO2

8 x 10-3

1

22

25

27

CO

2 x 10-3

1

54

64

75

N2

9 x 10-3

1

10.8

12.8

15

H2

-5

0.5

2.3

6.2

13

5 x 10

process (Equation (1)), most of the gas is given off during the initial period of operation of a VIP. Results shown in Figure 4.4 and Table 4.2 can vary from sample to sample depending on the open cell PU foam preparation technique, its microstructure and density and the pre-treatment. Since desorption is a thermally activated process [18], the outgassing rates increase as the temperature increases. The outgassing contribution has therefore to be carefully evaluated in all those applications where the vacuum panels operate continuously at temperatures higher than room temperature, e.g., 60-80 °C, or have to withstand high temperature peaks, for example 100 °C, even for a relatively short period of time. Examples of such applications are presented and discussed in Section 4.6. An additional factor which may have an impact on the vacuum properties is the closed cell content of the PU foam. All the open cell PU foams so far presented in the literature for VIP applications are quoted to be 100% open cell based on pycnometric measurement. On the other hand, the accuracy of this test method is generally close to ± 0.2% [15], so that the possibility of a small fraction of closed cells cannot be completely ruled out and has to be considered as an additional potential source of pressure build-up in the panel. Closed cells might be present due to a non optimal foam preparation or to the incomplete skin removal.

4.3.2 Vacuum Properties of the Barrier Film The barrier film plays an important role in vacuum panel technology since it has the task of minimising air and moisture penetration into the vacuum core. The barrier must be

167

Advances in Urethane Science and Technology durable and able to resist to puncturing and abrasion. It must be functional over a wide temperature range and retain its physical properties, such as dimensional stability, flexibility and sealability for years. Thermal conductivity of the skin should also be very low to minimise the heat transfer through the panel edges which can partially reduce the insulation efficiency of the evacuated panel (‘edge effect’, see Section 3.2.1). These properties have to be coupled with very low gas permeation and outgassing rates. Gas permeation through the barrier envelope is one of the most important factors responsible for the pressure increase in a panel during its life. Depending on the structure of the barrier film and the materials used, gas ingress can preferentially take place through the whole surface of the barrier film (permeation through the surface) or through the heat-sealed plastic VIP flanges (permeation at the edges), or both.

4.3.2.1 Gas Permeation Through the Surface Several types of barrier materials, having different structures and gas transmission rates, are commercially available from the food, packaging and electronic industry. Gas barrier requirements for vacuum panels are however much more demanding. Composite plastic films, obtained by laminating several polymeric sheets have been proposed in the past for precipitated silica-filled panels, where pressures as high as 5001000 Pa can be tolerated without excessive degradation of the thermal conductivity. In the case of open cell PU-filled panels, requiring much lower pressure values, this barrier structure is generally inadequate, if a lifetime exceeding one to two years is needed. To improve the permeation properties, barrier materials, such as aluminium and silicon oxide, can be vacuum deposited on the polymer surface in the form of a thin film (0.010.05 μm thick) [19]. However, most commercially available metallised samples do not provide enough barrier properties for long term applications, mainly due to water transmission through the large density of pin holes and micro-defects present in the sputtered aluminium layers [20]. To overcome this problem, barrier films obtained by laminating a thicker continuous metal foil with various polymeric sheets, such as polyethylene terephthalate (PET), Nylon and polyester (PE) have been proposed to the appliance industry some years ago and are now commercially available [15, 21]. Aluminium is the preferred choice for

168

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… the metal, mainly due to its good workability, which allows production and lamination of virtually pin hole-free foils having thickness of 6 μm or less [21]. The presence of the aluminium foil dramatically improves the gas barrier property of the film, well beyond the sensitivity limits of the analytical techniques commercially available for the measurement of gas permeation. In the case of water, which is one of the most important permeating species, two standard methodologies are at present widely used to measure its transmission rate through a material, as described by ASTM E96-00 [22]. According to the first procedure (‘the desiccant method’) the barrier sample is sealed to the open mouth of a test plate, which is cylindrical in shape, in which the desiccant is placed, and the assembly kept in a humidity and temperature controlled environment. The permeation of water through the sample is measured by the weight increase of the assembly. In the second procedure (‘water method’) the test plate contains water and the transmission rate through the barrier sample into the controlled environment is measured by the weight decrease of the assembly. A variation of the first method, frequently used in practice, is based on periodically weighing vacuum panels, containing of desiccant, and kept in a given temperature and humidity-controlled environment. Another very common technique is based on the use of an infrared sensor as described by ASTM F1249-90 [23]. The barrier sample is sealed between a dry and a wet chamber kept at known temperature and relative humidity. The two chambers make up a diffusion cell which is placed in a test station where the dry chamber and the top of the barrier are flushed with dry air. Water vapour which penetrates through the barrier film blends with the dry air and is transported to a pressure-modulated infrared sensor. The infrared radiation is absorbed by the water molecules and the sensor produces an electrical signal which is proportional to the concentration of water in the gas phase. The intensity of the signal is then compared with the signal generated by measurement of a sample having a known water transmission rate. This allows the determination of the water transmission rate through the test specimen. The detection limit for water transmission generally quoted for commercially available instruments based on this principle is in the 0.01 g/m2 day range (at 90% relative humidity and 23 °C). These techniques are generally more than adequate for many plastic or standard quality metallised films but can hardly discriminate between samples containing aluminium foil, which generally have water transmission rates much lower than this value. To better estimate gas permeation for high quality barriers a novel technique has been developed [24, 25], which is based on the measurement with a quadrupole mass spectrometer of the helium transmission rate through the sample.

169

Advances in Urethane Science and Technology

Figure 4.5 Scheme of the experimental bench for the helium permeation test.

A sketch of the experimental bench is shown in Figure 4.5. A picture of the bench is shown in Figure 4.6. The bench is a stainless steel ultra-high vacuum apparatus equipped with rotary, turbomolecular and non-evaporable getter (NEG) pumps. The sample, a circular coupon of about 30 cm2 area, is mounted between two flanges and the helium pressure is applied (generally in the 0.0001-0.1 MPa range) on one side of the sample, the other side being in view of the quadrupole mass spectrometer. A variable conductance C (l/s) is mounted between the mass spectrometer and the evacuation group, so that the helium flow rate F (Pa-l/s) through the conductance can be measured with the mass spectrometer according to the equation:

170

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.6 High vacuum bench for helium permeation tests (Reproduced by courtesy of SAES Getters SpA, Italy).

F = C (P1 - P2)

(7)

where P1 and P2 are the pressures before and after the conductance C, respectively. Under equilibrium conditions, which can be reached after some minutes or some hours, depending on the nature of the sample, the gas flow, F, provides the helium transmission rate. A typical graph showing the achievement of steady state conditions is given in Figure 4.7.

171

Advances in Urethane Science and Technology

Figure 4.7 Achievement of the equilibrium condition in a helium permeation test for a laminated film composed of 15 μm Nylon, 12 μm PET, 6 μm aluminium, 50 μm high density polyethylene (HDPE) (or Nylon 15 μm/PET 12 μm/Al 6 μm/HDPE 50 μm).

The use of the mass spectrometer ensures very high sensitivity for helium, better than 10-11 Pa-l/(s m2 Pa), mainly due to the negligible presence of helium as a constituent of the gas background. Due to its high sensitivity and the relatively short measuring time, this technique can be used effectively to support the development of improved laminates and also for quality control in a production environment. The estimation of the transmission rates for gases other than helium can also be obtained, provided a preliminary calibration is carried out. This is particularly important for water, which is the main gas permeating through the barrier skin. To run the calibration procedure, the helium transmission rate is measured for various samples having different and known permeation rates for water. The linear correlation between helium and water transmission rates is then established, as shown in Figure 4.8. The correlation factor for water was found to be close to 500 for most of the PE/PET/ Nylon-based films analysed by the authors, either metallised or incorporating an aluminium foil [26]. It was then possible to estimate the order of magnitude of the water

172

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.8 Experimental determination of the correlation factor between helium and water transmission rates.

transmission rate for an unknown low permeation rate sample by simply measuring the helium transmission rate and multiplying it by 500. Table 4.3 provides a comparison between the helium permeation rates in two films, a laminate incorporating a 6 μm aluminium foil, PET 12 μm/Aluminium 6 μm/HDPE 50 μm (Film B) and a multi-layered barrier composed of four aluminium-sputtered PET sheets laminated onto a 50 μm PE (Film A). Tests were run at 24 °C. From the measurement of

Table 4.3 Helium and water permeation rates through films Sample

Permeation rate (MPa s-1cm-2 Pa-1)

Estimated water amount in a VIP after 15 years

helium

water

Film A

910 x 10-12

460 x 10-9

100 g

Film B

-15

-13

0.02 g

130 x 10

670 x 10

173

Advances in Urethane Science and Technology

Table 4.4 Thermal conductivity of some materials used as gas barriers and/or structural layers in barrier films. Material

Thermal conductivity (μ /m-K)

Typical thickness (μm)

PET

0.15

10-15

HDPE

1.05

40-60

PVDC

0.13

10-20

Nylon

0.43

10-20

Aluminium

273

6-7

PVDC: Polyvinylidene chloride

the permeation rates, the quantity of water permeating in 15 years in a 50 x 50 x 2 cm3 vacuum panel has been estimated. The metallised barrier here considered cannot be used for long term applications, where 10-15 years or more are targeted, since the amount of desiccant necessary to compensate for the water transmission would be too large. A much lower permeation rate is provided by Film B, which is therefore well suited for long term applications. In spite of the extremely good gas barrier properties, VIP prepared with such a film suffers from an intrinsic limitation, i.e., the high thermal conductivity of the aluminium foil, as shown in Table 4.4. A fraction of the heat flow is, in fact, transferred from the hot to the cold panel surface by the aluminium foil through the panel flanges, rather than through the core material (so called ‘edge effect’). As a result, the average insulation value of the panel is less than the expected value based on the actual insulating properties of the core material (centre of the panel thermal conductivity), this difference being more remarkable the smaller the panel size. The edge effect can be evaluated, as a function of the panel size, the laminate structure and the aluminium thickness, by numerical methods, such as the Finite Element Analysis (FEA). Results of such an analysis are given in one specific example in Figure 4.9, where the average thermal conductivity of a vacuum panel using a PET 12 μm/Aluminium 6 μm/ HDPE 50 μm laminate is plotted against the panel area.

174

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.9 Thermal conductivity versus panel area.

The thermal conductivity and the thickness of the PU core material considered here are 6 mW/m-K and 2 cm, respectively. Figure 4.9 shows that the panel thermal conductivity approaches the value of the core material only for a sufficiently large panel area (≈ 1-10 m2). For very small panels, the insulating properties of the core material are spoiled by the edge effect and the energy savings in the real application may be minimal. For this reason, efforts are now being made to improve the aluminium–based laminates and/or to develop products not containing the aluminium foil but still having sufficiently good gas barrier properties, so as to achieve an acceptable trade-off between energy saving performances and acceptable lifetime. This challenging objective can be addressed by selecting and properly combining together suitable polymers having improved gas barrier properties. A typical example of this process, recently reported by Lamb and Zeiler [27], is given in Table 4.5, which shows how significant improvements can be obtained in non aluminiumfoil containing skins working on the film structure.

175

Advances in Urethane Science and Technology

Table 4.5 Oxygen transmission (OTR) and water (WTR) transmission rates for some film structures [27] OTR x 10-10 m2 day (STP)

WTR g/m2 day

PET

34.8

43.4

PET + PVDC

2.3

9.3

PET + PVDC + metallised layer

0.04

0.62

0.00015

< 0.15

Film structure

Mylar 200RSBL300 STP: Standard, temperature, pressure

A comparison of some different types of improved metallised and aluminium foilcontaining barriers, also including a preliminary investigation of the effect of temperature on the gas transmission rates, has been recently reported by Bonekamp [28]. Due to the still developmental nature of most of these barriers, more work is necessary to assess their usability for long term applications. Extensive characterisation of the mechanical, sealability and stress resistance properties are also important issues to consider.

4.3.2.2 Gas Permeation Through the Flanges The permeation of atmospheric gases through the polymeric sealed flange of a vacuum panel can be an important contribution to the pressure increase in a VIP for long term applications. The gas permeation rate depends on the pressure gradient across the vacuum panel, the type of polymer used as a sealant, the flange geometrical parameters (exposed surface and width) and temperature. Polyethylene is at present one of the preferred sealing materials due to the good trade-off achieved among sealability, mechanical properties, gas permeation, outgassing rates, reliability and cost. Gas permeability values quoted in the literature for PE are scattered over an order of magnitude, depending on the actual density of the sample and the gas considered [29, 30, 31]. Typical permeability values corresponding to medium and high density PE are shown in Table 4.6. Results of the calculation for argon, nitrogen, oxygen and water and the total pressure, are given in Figure 4.10 and show the role of permeation in deteriorating the vacuum level in the panel. The width and height of the PE flange here considered are 10 and 0.1 mm, respectively and the environment is air at 21 °C and 50% relative humidity.

176

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Table 4.6. Permeation rates for gases (cm3/m2d, STP) and water (g/m2d) in a 25 μm thick PE sheet Gas

MDPE x 10-10

HDPE x 10-10

Temperature °C

N2

387

96.75

21

O2

1548

425.7

21

Ar

3096

774

21

H2O

10.85

4.65

38, 90% RH

MDPE: medium density PE; RH: relative humidity

Figure 4.10 Pressure increase due to gas permeation at the panel PE flanges.

In the case of panels which are encapsulated, i.e., surrounded by a closed cell PU foam, as happens for panels in household refrigerators, freezers or vending machines (see Section 4.6), the picture is more complicated. The gas environment surrounding the vacuum panel will depend on the type of encapsulating foam and will change with time as a consequence of the gas out-diffusion from the closed cells and the progressive air penetration from the outside, thus making predictions more difficult. This ageing process

177

Advances in Urethane Science and Technology will also depend on the temperature and the hermeticity of the case surrounding the encapsulating foam as shown in various papers [32, 33]. Permeation of blowing agents through the seals, even though not a major contribution, has also to be considered. The use of a different PE grade or a different sealant polymer will provide different pressure build-up curves in the panel both in terms of total and partial gas pressures. High barrier polymers, such as polyvinylidene chloride (PVDC), polyacrylonitrile (PAN), polyester, acid copolymers and ionomers have been proposed and are under evaluation as a replacement for PE. However, they present some drawbacks in terms of sealability and/or mechanical properties and/or outgassing rates. Cost of these polymers is also generally higher.

4.3.2.3 Outgassing Properties of the Film In general, very little data have been published on the outgassing properties of the skins for vacuum panels, even though they can contribute, in some cases, in a non-negligible way to the deterioration of the pressure inside the VIP. This can be due to the outgassing properties of the materials used as barrier layers and/or the lamination process, which may introduce volatile substances or trap gases in between the various sheets. The outgassing properties of a barrier sample can be measured using the same experimental equipment shown in Figure 4.2 and have to be taken into account to estimate the total gas load in the panel. For the same film having the barrier property in Table 4.3, the extrapolation of the outgassing experimental data to 20 years in a 50 x 50 x 2 cm3 size panel is shown in Table 4.7. Water, which is also given off by the barrier, was not quantified in this test.

Table 4.7 Outgassing parameters and gas load for a barrier incorporating a 6 μm aluminium foil Gas

q0 (Pa-l/h)

α

Q (Pa-l)

H2

35 x 10-4

0.5

29

N2

-1

1.3

133

178

43 x 10

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

4.3.3 The Getter Device As discussed in the previous sections, gas desorption from the surfaces exposed to the vacuum (filler and barrier) and gas permeation through the bag contribute to increase the pressure in a VIP during its life. These contributions, as well as the residual gases left in the panel after the exhaust and seal-off process, have to be taken into account to provide an accurate estimation of the pressure increase in a VIP as a function of time. Figure 4.11 shows the total pressure increase, not including water, for some panel sizes, when data from Tables 4.2, 4.3, 4.6 and 4.7 are used.

Figure 4.11 Estimated total pressure increase in some open cell foam-filled VIP as a function of time. The additional pressure increase due to water vapour is not considered in this calculation.

179

Advances in Urethane Science and Technology For given temperature and environmental conditions, the pressure build-up depends on the specific size of the panel, and in particular on the ratio between the perimeter and the thickness, the lower the latter the higher the pressure increase. A barrier film incorporating an aluminium foil and a panel seal-off pressure of 5 Pa have been here considered. In this example, water has been assumed to be completely absorbed by a proper amount of desiccant and not to contribute to the pressure increase. In the case of a typical 50 x 50 x 2 cm3 size VIP, the pressure build-up exceeds 100 Pa after a few years and even not considering water, the total pressure after 20 years exceeds 250 Pa. The deterioration of the thermal conductivity for such a panel, a direct consequence of the increase of the internal pressure, is shown in Figure 4.12, for a PU foam following the curve of Figure 4.1. Due to the nature of the thermal conductivity versus pressure curve, the deterioration of the λ factor is negligible at the beginning but increases steadily with time.

Figure 4.12 Increase of the thermal conductivity in a vacuum following the pressure build-up as per Figure 4.11. Water is supposed to be absorbed by a desiccant and not contribute to the pressure rise.

180

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… In some applications, such as in domestic refrigerators and freezers, an additional gas burden may be also generated in the panel during the appliance manufacturing process or by the appliance operating conditions, which may even accelerate the deterioration of the thermal conductivity. These examples show that for medium and long term applications (> 2-3 years) a simple desiccant is not enough and specific getter materials have to be added to sorb the extra amount of gases generated in the panel. The wide spectrum of gases present in a panel, including carbon dioxide, carbon monoxide, hydrogen, nitrogen, oxygen and water, as well as traces of blowing agents, also calls for an absorption system having high sorption capacities and adequate pumping speed for all these gases. Zirconium-based NEG alloys, which are widely used in a variety of vacuum applications [34, 35, 36, 37] cannot be used in plastic evacuated panels due to their limited sorption capacity at room temperature and the need to be heat activated at relatively high temperature (> 350 °C) prior to their use, this process being clearly not compatible with the panel polymeric components. Very large area physical adsorbents, like molecular sieves, zeolite or activated charcoal [38, 39] have very good efficiency for water and some organics but present serious limitations in sorbing carbon monoxide, hydrogen, nitrogen and oxygen at the temperature and pressure conditions typically encountered in VIP applications. Therefore, they have beneficial effects during the very initial life of the VIP without being able to ensure long lifetime, as required, for example, by the appliance industry. They are also sensitive to the sorption temperature, i.e., the higher the temperature the lower the sorption performance, which spoils sorption performances in high temperature applications, e.g., 60-100 °C. Physical adsorbents have also a second drawback which is related to their treatment, when mass production quantities have to be handled. To fully take advantage of their sorption capacity, in fact, a relatively high temperature pre-treatment process which cleans the surface by promoting gas desorption is required. However, after this treatment, the adsorbents are exposed again to the ambient environmental conditions during the subsequent handling and mounting operations. Due to their reactivity, they will rapidly sorb some gases from the environment and this will reduce their capacity in a poorly reproducible way, critically dependent on the environmental conditions, temperature and exposure time. Fluctuations in their sorption performances as well as in the VIP insulating efficiency could be the final result. To eliminate the above problems and address the lifetime issues posed by the appliance industry, a novel getter device, the COMBOGETTER, has been recently proposed [40]. 181

Advances in Urethane Science and Technology At the heart of this device is a barium-lithium alloy, in a 1 to 4 atomic ratio [41], able to efficiently chemically absorb a large amount of nitrogen at room temperature, up to more than 2500 Pa-l (N2)/g (alloy) [42]. This very specific feature allows the getter to compensate for the air inlet coming from permeation, thus ensuring long lifetime requirements. High efficiency calcium oxide and a metal oxide are also added to barium-lithium to absorb moisture, hydrogen and some of the most common blowing agents, such as R141 b and cyclopentane which could permeate through the VIP during its life. All these materials are prepared according to proprietary processes, which confer upon them unique properties in terms of sorption efficiency and capacity. The active powders are compressed in a stainless steel cup according to a configuration which allows optimum sorption performances and ease of use (Figure 4.13). The getter device does not need to be heat activated or pre-treated before being used in VIP and can be handled in air for a reasonable period of time (several minutes), during the panel manufacturing process, without affecting its sorption capacity. Mounting can be accomplished by inserting the getter in a recess cut in the filling material. Alternatively, if the filler is sufficiently soft, as is the case for open cell PU, the getter can be simply pressed into it.

Figure 4.13 Picture of the COMBOGETTER device for VIP.

182

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Table 4.8 Gas sorption capacity of the COMBOGETTER Gas type

Gas capacity, Pa-1

Air

530

CO/CO2

800

H2

9300

H2O

80000

Blowing Agents

130

Gas capacity data for the getter are summarised in Table 4.8. A typical sorption curve for nitrogen at room temperature is given in Figure 4.14 (sorption throughput, Pa-l/s, as a function of the sorbed amount, Pa-l).

Figure 4.14 Sorption curve for nitrogen for the COMBOGETTER (at room temperature)

183

Advances in Urethane Science and Technology

Figure 4.15 Sorption curve for nitrogen for the COMBOGETTER (in the 20-70 °C range)

This throughput is adequate to compensate for the relatively slow gas inlet rate occurring during the VIP lifetime. The effect of the temperature on the sorption performance is shown in Figure 4.15 in the range 20-70 °C. These data show that the getter is effective also at relatively high temperatures, where physical adsorbents become less efficient. The getter in VIP technology has several roles, as demonstrated by the following examples: a) Life insurance Permeating and outgassed species (carbon dioxide, carbon monoxide, hydrogen, nitrogen, oxygen, water) which can significantly increase the pressure are irreversibly fixed by the getter, as shown by tests of actual panels (Figure 4.16). Pressure in VIP has been recorded by means of a viscous pressure gauge, as explained in Section 4.4.2.2. The resulting thermal conductivity of gettered panels is therefore much lower in the long term, as shown in Figure 4.17, which compares the λ factor in gettered and ungettered panels after almost 3 years of ageing.

184

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.16 Measurement of the pressure in two vacuum panels (with and without getter)

Figure 4.17 Measurement of the thermal conductivity in two panels (with and without COMBOGETTER) aged at room temperature

185

Advances in Urethane Science and Technology b) VIP manufacturing process The ability of the getter to sorb air can be also used to reduce the evacuation time in panels, this contribution being particularly interesting for medium and large size panels, e.g., > 0.5 m2, where achieving base pressure in the low 1 Pa can be time consuming for open cell foams with very small cell openings, which limit the evacuation efficiency. For each application, a trade-off between the getter capacity needed to reduce the evacuation time and that necessary to ensure long lasting service performance has to be found. A typical example of the getter’s ability to act as a chemical ‘in situ’ pump to sorb residual gases after the evacuation process is shown in Figure 4.18, which compares the pressure trend in panels with and without the getter immediately after the evacuation and seal-off.

Figure 4.18 Pressure decrease in a VIP due to the in situ pumping effect of the COMBOGETTER.

186

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… c) Process yield and heavy duty conditions The COMBOGETTER normalises the vacuum level in the production cycle by compensating for fluctuations in the quality of the VIP components and the manufacturing process. It can also compensate for the gases generated in the panel during ageing or heat treatment, as happens during storage in the warehouse, the foaming process in a refrigerator cabinet or the high temperature operating conditions required by several specific applications. Experimental data on vacuum panel performances under these conditions are presented and discussed in Section 4.6. For some specific VIP sizes and short term applications (for example one to two years) a getter for air may not be necessary, since the permeation contribution is less relevant here. In these cases water is usually the most important gas to be sorbed and the use of a simple dryer is sufficient to keep the pressure at the required 10 Pa value. Calcium oxide is one of the preferred desiccants due to its availability, low cost, good environmental features and large water sorption capacity. The production process influences the calcium oxide physical structure (particle size, porosity and morphology) and thus the water sorption efficiency may be actually an important parameter to consider in short-term applications (for example in shipping containers, where lifetimes as short as a few months or less, can be of interest). A comparison between a commercially available product and a highly efficient calcium oxide, prepared by SAES according to a proprietary process is given in Figure 4.19. Also physical absorbents, like molecular sieves and silica gel or activated carbon, can be used for short-term applications.

187

Advances in Urethane Science and Technology

Figure 4.19 Water sorption test on two calcium oxide samples. Samples were exposed at 100-200 Pa water partial pressure at 23 °C. Sample A was prepared by SAES according to a proprietary process to increase material surface area and porosity.

4.4 Vacuum Panel Manufacturing Process and Characterisation 4.4.1 Some Manufacturing Issues Before being inserted in the barrier bag, the open cell foam needs preliminary pretreatment in air to remove water and the residue of the foam production process. Volatile materials are in fact used to blow the foam chemicals and remain partially trapped in the foam matrix or condense as a liquid on the foam surface after the cells have been opened. Even though part of these volatile substances will desorb at room temperature, the remaining quantity will generate an unacceptable pressure build-up in the panel [43]. The baking process is also necessary to remove the water absorbed by the PU foam when exposed to humid air (about 1.8 % by weight, at 23 °C, 50% relative humidity), this being particularly effective to reduce the evacuation time, as shown by tests of actual panels [44]. Pre-treatments generally range from 150 °C for 10 minutes to 120 °C for up to a few hours [9, 12, 43], depending on the foam type.

188

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.20 Water re-absorption curve of a small PU sample after baking.

After the foam has been thoroughly baked, it is important to reduce its exposure time to the working environment to avoid significant moisture re-absorption, which takes place when PU foam cools down. As an example, the weight increase due to water re-absorption during cooling to room temperature is shown in Figure 4.20 for a small PU sample, dried at 150 °C for 20 minutes and then exposed to 50% relative humidity. These data indicate the need to design the production process and equipment in such a way that the foam is exposed to air for a very short time, not exceeding a few minutes. A limited exposure is also beneficial to reduce the evacuation time of the panel, which strongly depends on the desorption rate of physisorbed water. To further minimise reabsorption of water on the foam it is advisable to process the PU in a dry or humiditycontrolled area. The water content picked-up by the foam can be measured in the finished panel by means of residual gas analysis (RGA) with the mass spectrometer, as described in Section 4.4.2.2. In general, the proper pre-treatment of the foam followed by adequate handling ensures better initial insulation performance and shorter evacuation cycles which also mean higher productivity and lower costs per panel. The open cell foam slabs are generally easier to handle and process than powder-based core materials like precipitated silica or perlite. However, precautions have to be adopted

189

Advances in Urethane Science and Technology during handling of the foam to avoid powder generation which could have an impact on the quality and reliability of the seals. Care also has to be taken during the production process to ensure adequate VIP flatness, which is an important factor for the subsequent proper mounting of the panel in various applications. Appropriate film handling and processing is also the key to produce reliable and good quality panels. Gas barrier properties of films, in fact, mainly depend on the number of defects, such as cracks and pinholes, present in the layered structure. This number increases as a consequence of the mechanical stresses the film undergoes during its handling. Moreover, the high load applied by the atmospheric pressure on the vacuum panel, this value being particularly high in the corners and along the edges, stretches the layers and locally increases the defect density. A preliminary investigation of these effects has been carried out by Sugiyama and coauthors [21] for some laminated plastic films incorporating a 6 μm aluminium foil. Films have been analysed as such and after having been twisted to simulate handling conditions. VIP have also been prepared and samples cut from the flat surface of the panel, the edges and the corners to check for any local increase of the permeation rate. Tests were carried out at 23 °C on samples of 2.5 x 10-3 m2 area applying 0.0505 MPa helium pressure. The structure of the analysed films was Nylon 15 μm/vacuum metallised PET 12 μm/aluminium 6 μm/HDPE 50 μm. The results, shown in Table 4.9 clearly indicate that the twisting process is responsible for significantly larger permeation rates. Samples cut at the corners also show higher permeation rates, which confirms that stress conditions in these areas increase the local pin-hole density in the aluminium foil.

Table 4.9 Helium permeation rates through barrier films submitted to different mechanical treatments Sample number

Type of sample

He permeation rate (Pa-l/s) x 10-11

1

Reference film (no treatment)

0.5 m per side. If the panel is encapsulated, as actually happens in several applications, the conventional foam surrounding it also has to be taken into account in the ECR calculation, since it may contribute to heat dissipations, its effect being particularly important around the VIP flanges. The ECR values for encapsulated panels of different sizes prepared with a laminate incorporating a 6 μm aluminium foil (Table 4.3, Film B) are given in Table 4.10. The calculation has been made assuming a foam thermal conductivity of 6 mW/m K and a thickness of 0.5 cm of additional PU foam (thermal conductivity of 20 mW/m K) for a total thickness of 0.025 m.

197

Advances in Urethane Science and Technology

Table 4.10 ECR values in encapsulated VIP of different sizes VIP size (cm)

ECR (%)

40 x 40 x 2

16.5

50 x 50 x 2

26.7

50 x 100 x 2

36.7

100 x 100 x 2

46.1

The ECR values achievable with the aluminium foil-based laminates are very interesting for large size panels, where the edge effect is not severe. For small panels, which are also important, since they can be easily adopted in a variety of applications, from household appliances to shipping containers, the improvement in the insulation efficiency is not as remarkable. The energy consumption reduction in two refrigerator cabinets, each using one panel, 100 x 50 x 2 cm3 large, covering 40% and 60% of the total surface is shown in Table 4.11. The same assumptions used in Table 4.10 have been made for the PU thermal conductivity and thickness of the barrier foil.

Table 4.11 ECR values for two refrigerators with varying VIP coverage VIP size (cm)

Refrigerator/freezer surface coverage

ECR

100 x 50 x 2

60%

22%

100 x 50 x 2

40%

14.7%

In general, for household appliances, depending on the surface coverage of the cabinet and the panel thickness, energy savings from 10% to 30% have been reported using open cell PU foam-filled panels packaged in a 6 μm aluminium foil-based barrier [12, 15, 52, 56].

198

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… The influence of the vacuum panel thickness to the total insulation thickness ratio as well as the change in performance, by increasing the foam λ value, has been determined by Hamilton [54]. In general, results can be optimised by adjusting the VIP size and thickness to provide the most cost effective option. Further improvements can be obtained by decreasing the aluminium foil thickness or using a metal foil-free barrier and/or decreasing the thermal conductivity of the foam. The latter can be achieved by careful control of the foam microstructure and in particular the cell shape, orientation and size, which play an essential role in determining the foam thermal conductivity and its dependence on gas pressure. Conductivity values close to 5 mW/m K have been quoted for properly prepared open cell foam samples [9].

4.6 Examples of VIP Applications and Related Issues 4.6.1 Household Appliances Refrigerators and freezers account for about 20% of the total electricity consumption of household appliances. For this reason the appliance industry is under pressure to improve the energy efficiency of their products to cope with the need to reduce carbon dioxide emissions, as recently mandated by the Kyoto Conference. This objective has to be achieved without penalising product performance. Several options to decrease the energy consumption are under evaluation, ranging from high efficiency compressors to the adoption of intelligent electronic devices [57, 58, 59]. Improvement of the insulation through the use of vacuum insulated panels filled with silica powder or glass wool as core materials has been evaluated, and adopted in limited amounts, by most refrigerator manufacturers [60, 61, 62], cost being the main obstacle to large sales of these products. The lower cost of the open cell foams, combined with the low density and good handling properties, has generated renewed interest for this technology as an environmentally friendly option to energy consumption reduction. A second driving force for the appliance industry is the possibility of increasing the internal volume of refrigerators and freezers without increasing the outer dimensions. This aspect is particularly important in Europe and Japan, where the built-in appliance market is an important segment and space constraints play a role. Adoption of PU-based vacuum panels in refrigerators and freezers requires the proper handling of the foaming process, which may influence the structural and vacuum properties of the panel.

199

Advances in Urethane Science and Technology In fact, during the appliance manufacturing cycle, the panels are glued to the cabinet walls and then foamed in place with a conventional closed cell PU foam, this process being necessary to further improve the overall structural and thermal insulating properties of the cabinet. To allow a better flowability of the injected chemicals, the cabinets are generally pre-heated at about 45 °C for some minutes. The presence of the panels, which reduces the free flow and expansion of the foam, requires special care to avoid generation of air-filled void volumes which would deteriorate the insulation performance [56]. To prevent this, and provide the best integration between VIP and conventional PU foams, specific systems have been recently developed to be used in conjunction with VIP [10]. As a combined effect of the cabinet pre-heating and the exothermicity of the foaming reaction, temperatures even higher than 100-120 °C can be reached on the VIP surface, leading to temperature-enhanced gas desorption. A typical result using a 50 x 50 x 2 cm3 panel is shown in Figure 4.25.

Figure 4.25 Measurement of the temperature increase experienced by the two panel surfaces during the foaming process.

200

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… This gas contribution has to be evaluated case by case since it depends strictly on the actual foaming conditions, such as the chemicals used, the process variables, the refrigerator design and VIP geometry. However, under some circumstances, the pressure increase can be a measurable fraction of the maximum acceptable level, thus causing a deterioration of the VIP thermal insulation properties from the very beginning of the refrigerator life. As an example, Figure 4.26 shows the effect of a thermal treatment carried out at 50, 70 and 90 °C on a VIP prepared with an open cell foam and an aluminium foil-containing laminate. The foam was pre-baked at 150 °C for 20 minutes and no absorbent was used. After the preparation, the panel was put in a oven and kept at the indicated temperature for 15 to 20 minutes. The panel was then removed and the total pressure measured as a function of time. To allow the continuous monitoring of the pressure, a SpiroTorr spinning rotor gauge was mounted in the panel as shown in Figure 4.21. Desorbed gases after the thermal treatments are only partially reabsorbed by the foam and generate a measurable pressure build up in the VIP.

Figure 4.26 pressure evolution in a panel submitted to various heat treatments. Pressure readings are taken by means of a SpiroTorr SRG (see Figure 4.3).

201

Advances in Urethane Science and Technology

Table 4.12 Residual Gas Analysis in encapsulated panels with and without the getter Encapsulated VIP

Total Pressure (Pa)

Partial Pressure (Pa) Air

CO/CO2

H2O

H2

Without getter

40

6

29

5

0.1

With getter

0.7

0.3

0.01

0.3

0.1

The effect of the temperature increase resulting from the foaming process, on the vacuum level in a panel can also be seen in Table 4.12 which shows the result of an RGA carried out on two test panels encapsulated by an appliance manufacturer. In the getterless panel, the pressure immediately after the foaming is already exceeding the 10 Pa target value and close to the maximum acceptable value (generally set at 50 Pa). The deterioration of the vacuum during the foaming process has been recently addressed by Kücükpinar and co-workers who ran specific tests aimed at quantifying this effect in open cell foam-filled VIP with and without getters [53]. As shown in this study, the COMBOGETTER was able to compensate for the extra gas load generated during the foaming process, thus confirming the results in Table 4.12. Further specific outgassing studies are necessary to understand the mechanisms of gas generation during the foaming process, even though this investigation presents several difficulties because of the large number of parameters involved. The foaming process may also stress the laminate due to the combined effect of temperature and applied pressure on critical areas such as the panel edges and the corners. Defects generated here may in turn increase the gas diffusion inside the panel, reducing the designed lifetime. Another important issue in refrigerator/freezer applications is longevity, since 15 to 20 years lifetime is targeted by most manufacturers. Durability tests on panels are therefore necessary to assess VIP reliability and performance over time. Figure 4.27 shows the pressure values in some encapsulated panels, as measured by RGA. Panel A, containing a COMBOGETTER, was aged at room temperature for some weeks and then kept at 40 °C for more than 30 months. Panels B and C did not have any adsorbent. The latter was aged at 40 °C for about 3 months, while the former was unaged.

202

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.27 Residual gas analysis in panels encapsulated and aged for different times. Panel A contains one COMBOGETTER, panels B and C are without.

The pressure in panel A is well below the 10 Pa target value as compared to panel C whose pressure is one order of magnitude higher. Pressure in panel B, which was just encapsulated, was slightly lower than in panel C. For comparison, the calculated pressure increase in the panel, as predicted based on the model discussed in Section 4.3, is also shown. This result confirms the role of the getter to compensate for the encapsulation process and to ensure VIP longevity and also supports the gas load model previously described. Still, a better assessment is necessary through longer term tests and, possibly, through accelerated tests able to provide experimental data within a shorter time. Activity is ongoing in SAES Getters Laboratories, as well as in various other research centres to provide such an assessment.

4.6.2 Laboratory and Biomedical Refrigerators This is very special equipment designed to operate at very low temperatures, e.g., from -30 °C to -86 °C, to age samples or to store valuable and perishable goods, like organs and tissues, biological and medical samples or vaccines. Vacuum panels are used mainly to increase the internal storage volume without increasing the energy consumption.

203

Advances in Urethane Science and Technology Since the conventional insulation must be very thick to ensure the achievement of such low temperatures, a partial replacement of the conventional insulator with VIP contributes significantly to increase the internal volume, from 20% up to 40% or more in specific models. As a second advantage, panels provide an overall superior passive insulation, which means that the temperature rise in the case of power failure is less steep and more time is necessary before a given critical temperature is reached. This provides extra-safety for delicate articles which may rapidly deteriorate when their temperature exceeds a given value. Due to the generally higher insulation thickness, the foaming process here may be even more exothermic than in domestic appliances. Means have therefore to be taken to keep the VIP temperature as low as possible during the encapsulation. Ultra-low temperature freezer models using foam filled vacuum panels were successfully placed on the market by a leading company some years ago and other companies are expected to follow soon.

4.6.3 Vending Machines Vending machines are especially popular in Japan, Korea and the Far East. It is estimated that in Japan about 2.5 million of these appliances are in the field with a replacement market of about 0.4 million pieces/year. VIP are finding widespread use in vending machines since most manufacturers are extensively using one or two panels to separate the hot and cold beverage compartments. Also in this case, the main driving force for VIP adoption has been the possibility to increase the internal volume for the storage of beverages, rather than the improvement in energy efficiency. Vending machines are quite a demanding application for VIP, since the panel operating temperature is cycled between room temperature and 60-70 °C and the lifetime is 5 years. As already discussed, the high temperature promotes higher diffusion and outgassing rates and provides additional mechanical stress to the envelope. Ageing tests have been carried out by vending machines manufacturers to assess VIP usability and reliability in this application. Results obtained in SAES Getters Laboratories for 50 x 50 x 2 cm3 panels aged for several months at 60 °C are shown in Figure 4.28.

204

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels…

Figure 4.28 Measurement of the thermal conductivity as a function of time for panels aged at 23 and 60 °C, with and without the COMBOGETTER.

Again, the deterioration of the thermal conductivity, due to the pressure increase, is quite dramatic in the getterless panels both at 23 and 60 °C. As expected, the thermal conductivity increase is much more rapid for the panels aged at 60 °C than for the panel kept at room temperature. Only minimal differences are measured in the panels containing the COMBOGETTER, which could compensate for the increased gas load at both temperatures.

4.6.4 Refrigerated/Insulated Transportation Large vacuum insulated containers, ships and trucks are under evaluation by various companies, mainly in Europe and Japan. The main advantages offered by using VIP are space saving and better insulation, so as to keep the temperature increase rate as low as possible even in the absence of a power source. The lifespan is generally from 10 to 15 years. Since very large VIP, e.g., 2-3 m2 size are considered for this application, the edge effect (see Section 4.3.2.1) is here a marginal issue.

205

Advances in Urethane Science and Technology Small commercial shipping containers to store and deliver pharmaceuticals, frozen food and valuable temperature-sensitive products is also another interesting area for the use of VIP. Long lifespan is not a problem, since most of these shipping containers have a projected life of the order of weeks or months. Therefore, it is even possible to use the VIP more than once, the specifications on the permeation properties of the barrier film are less tight and the getter not necessary. A specific dryer, to adsorb water, may be used in most cases. The small size VIP and the limited projected life suggest the possibility of using metallised barriers for the skin of the panel in these applications, at least on one side, so as to overcome the edge effect, which, in this case is very critical.

4.6.5 Other Applications A variety of other potential applications exist for VIP, from cold stores to insulation in buildings and in industrial plants, e.g., industrial reactors or liquid natural gas tanks. Also super insulation at relatively high temperature applications for water heaters, heat pipes or boiling pots have been considered and are under evaluation. The key issue of the adoption of VIP in this last family of applications is the possibility to bend the open cell PU filled VIP into a round or cylindrical shape. The combination of high temperature and bending provides additional challenges to VIP technology and further improvements in the material selection and processing might be required.

4.7 Near Term Perspectives and Conclusions The open cell PU foam is a very promising filler material for VIP. It shows good vacuum compatibility, thermal conductivity, light weight, ease of handling and processing and moderate cost. To allow its use, significant improvements have been made recently in the barrier and getter technology, as well as in the panel manufacturing and testing procedures. Laminated barriers, incorporating a 6 μm continuous aluminium foil, are providing a reasonable trade-off between permeation properties and energy saving, especially in large size panels (~1 m2), where the edge effect is less important. These films can therefore be used in applications such as refrigerators, freezers, vending machines and insulated transportation, where 10-20 years lifetime and relatively large panels are required. Traditional metallised or multilayered plastic barriers, even though more appealing from the energy saving point of view, are not advisable in long term applications, due to their poorer gas barrier properties. However, the need to improve performance and VIP cost effectiveness is pushing the film industry to further improve their products either by

206

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… reducing the aluminium foil thickness to 5 μm and below or developing aluminium-free composite barriers. If successful, this latter solution would dramatically increase the vacuum panel insulation efficiency. New families of improved metallised products have been developed very recently and proposed for testing and are under evaluation in several laboratories. The outgassing and permeation data, on both components and finished panels, and the use of models quoted in the literature, allow the evaluation of the total gas load and the pressure increase in a vacuum panel as a function of time, size and operating conditions. This in turn is beneficial for designing the most suitable solution for the absorbent. A high capacity getter system, the COMBOGETTER, capable of adsorbing air, moisture and the other gases of interest without the need for any pre-treatment has been specifically developed for VIP application. Its roles are many since it chemically absorbs residual, outgassed and permeating species and provides the means to shorten the evacuation process and increase the manufacturing yield. Mass production technologies to manufacture high quality open cell foam panels are on the market. Several techniques are also becoming available for the fast measurement of the thermal conductivity. This is a very important issue to build a QA/QC system able to ensure the quality and the reliability of the products, not only after production but also immediately before use. As far as the applications are concerned, open cell PU foams are finding their place in the market for both low and high temperature appliances, from refrigerators to vending machines. Present production volumes are still limited but near term perspectives are encouraging, especially in Asia and Europe. The strong push for energy reduction and the parallel continuous refinement of the technology should provide further motivation to VIP adoption. Cost is still the main obstacle to the widespread adoption of this technology. A preliminary qualitative cost comparison between silica and open cell PU, which includes raw material, panel manufacture and panel installation costs [10] shows that a cost reduction of about one-third over silica could be achieved. This has to be improved to bring the cost of the vacuum panel down further, so as to really interest more segments of the insulation industry. Very recent breakthroughs in the foam mass production technologies seem to indicate that this target is reachable. Even though far from being complete, the material presented in this chapter shows that significant technological achievements have been obtained during the last few years.

207

Advances in Urethane Science and Technology Thanks to this, vacuum panel technology is becoming a technically viable and cost effective solution to the need to reduce energy consumption in household appliances and in commercial and industrial applications. To successfully achieve this target, additional efforts are necessary to further improve the component quality and reliability (foam, film and adsorbent), to optimise panel production and to reduce costs.

Acknowledgements The author would like to thank Dr. Paolo della Porta, President and CEO of SAES Getters SpA, for his continuous support and long lasting commitment to VIP technology. The author also acknowledges Dr. Bruno Ferrario, Corporate R&D Director, for fruitful discussions and suggestions. Special thanks are given to Dr. Roberto Caloi and Dr. Enea Rizzi for their valuable technical support in the development of theoretical models and in running experimental measurements on VIP components and finished devices.

References 1.

H. A. Fine, Journal of Thermal Insulation, 1989, 12, 183.

2.

D. W. Yarbrough, R. S. Graves, F. J. Weaver and D. L. McElroy, presented at ORNL/ CON-215, Oak Ridge National Laboratory, Oak Ridge, TN, Sept. 1986, p.1.

3.

L. R. Glicksman and N. Solomou, Presented at the SPI Polyurethanes Conference, Boston, MA, 1994, p.182.

4.

H. M. Strong, F. P. Bundy and H. P. Bovenkerk, Journal of Applied Physics, 1960, 31, 1, 39.

5.

B. Veltkamp and P. Hoogendoorn, Vuoto, 1999, 28, 1-2, 27.

6.

J. Fricke, Presented at the 13th European Conference on Thermophysical Properties, Lisboa, Portugal, 1993, p.207.

7.

R. L. Tabor, Presented at the SPI Polyurethanes Conference, Boston, MA, 1994, p.288.

8.

G. Biesmans, D. Randall, E. Francais and M. Perrut, Presented at the Polyurethanes World Congress ’97, Amsterdam, The Netherlands, 1997, p.294.

208

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… 9.

K. Kodama, K. Yuge, Y. Masuda and Y. Tanimoto, Presented at the Polyurethanes World Congress, October 1993, p.140.

10. R. De Vos and I. D. Rosbotham, Cellular Polymers, 1994, 13, 2, 147. 11. K. W. Dietrich and D. W. McCullogh, Presented at UTECH 96, The Hague, The Netherlands, 1996, Paper No.64. 12. W. H. Tao, W. F. Sung and J. Y. Lin, Journal of Cellular Plastics, 1997, 33, 3, 545. 13. P. Pendergast and B. Malone, Vuoto, 1999, 28, 1-2, 27. 14. L. R. Glicksman in Low Density Cellular Plastics: Physical Basis of Behaviour, Ed., N. C. Hilyard and A. Cunningham, Chapman & Hall, London, 1994, Chapter 5, p.104. 15. W. Wacker, A. Christfreund, D. Randall and N. W. Keane, Presented at the Polyurethanes Expo ’96, Las Vegas, NV, 1996, p.35. 16. A. Roth, Vacuum Technology, North Holland Publishing Company, Amsterdam, 1982, 186. 17. F. Sciuccati, G. Gasparini and B. Ferrario, Vuoto, 1988, 17, 4, 345. 18. P. A. Redhead, Vacuum, 1962, 12, 203. 19. P. Mercea, L. Muresan and V. Mecea, Journal of Membrane Science, 1985, 24, 297. 20. T. A. Beu and P. V. Mercea, Materials Chemistry and Physics, 1990, 26, 309. 21. A. Sugiyama, H. Tada and M. Yoshimoto, Vuoto, 1999, 28, 1-2, 51. 22. ASTM Standard E96-00 Standard Test Methods for Water Vapor Transmission of Materials. 23. ASTM Standard F 1249-90 (1995) Standard Test Method for Water Vapour Transmission Rate Through Plastic Film and Sheeting Using a Modulated Infrared Sensor. 24. P. Manini, inventor; Saes Getters SpA, assignee; WO 9803850A1, 1998. 25. P. Manini, Journal of Cellular Plastics, 1999, 35, 5, 403.

209

Advances in Urethane Science and Technology 26. P. Manini and E. Rizzi, Poster presented at the European Workshop on Vacuum Panel Technology for Super Insulation in Domestic Appliances and Industrial Applications, Milano, Italy, 1998. 27. W. Lamb and R. Zeiler, Presented at the Vacuum Insulation Panel Symposium, Baltimore, MD, 1999. Paper available from the SAES Getters web site, http:// www.saesgetters.com. 28. J. Bonekamp, Presented at the Vacuum Insulation Panel Symposium, Baltimore, MD, 1999. Paper available from the SAES Getters web site, http:// www.saesgetters.com. 29. R. Juran, J. Covington and J. Carley, Modern Plastics Encyclopedia, McGraw Hill, NewYork, 1987, 64, 10A. 30. Polymer Handbook, 3rd Edition, Ed., J. Brandrup and E. H. Immergut, John Wiley and Sons, New York, 1989. 31. S. J. Adam and C. E. David, Presented at the 23rd International SAMPE Technical Conference, New York, NY, 1991. 32. K. E. Wilkes, W. A. Gabbard and F. J. Weaver, ORNL/CP-95971, Oak Ridge National Laboratory, TN, 1998. 33. G. Biesmans, R. De Vos and I. D. Rosbotham, Presented at the Polyurethanes World Congress 1993, Vancouver, BC, 1993, 498. 34. A.Barosi in International Symposium on Residual Gases in Electron Tubes, Ed., T. A. Giorgi and P. della Porta, Academic Press, London, 1972, 221. 35. C. Boffito, B. Ferrario, P. della Porta and L. Rosai, Journal of Vacuum Science and Technology, 1981, 18, 1117. 36. B. Ferrario, A.Figini and M.Borghi, Vacuum, 1984, 35, 13. 37. C. Boffito, B. Ferrario, L. Rosai and F. Doni, Journal of Vacuum Science and Technology, 1987, A5, 3442. 38. D. W. Breck, Zeolite Molecular Sieves, R. E. Krieger Publishing Company, Malabar, Florida, 1984, Chapter 8. 39. J. S. Mattson and H. B. Mark, Activated Carbon: Surface Chemistry and Adsorption from Solution, Marcel Dekker, New York, 1971.

210

Recent Developments in Open Cell Polyurethane-Filled Vacuum Insulated Panels… 40. P. Manini, Vuoto, 1997, 27, 2, 45. 41. C. Boffito and A. Schiabel, inventors; Getters SpA, assignee; European Patent 0514348 A1, 1991. 42. R. M. Caloi, P. Manini, S. Valdrè, E. Magnano, J. Kovac, E. Narducci and M. Sancrotti, Journal of Vacuum Science and Technology, 1999, A17, 2696. 43. R. De Vos, I. D. Rosbotham and J. Deschaght, Presented at the Polyurethanes World Congress, 1994, Boston MA, p.194. 44. J. Akita, Vuoto, 1999, 28, 1-2, 59. 45. ASTM Standard C518-98, Standard Test Method for Steady-State Thermal Transmission Properties by Means of the Heat Flow Meter Apparatus. 46. ASTM Standard C236-89 (1993) e1, Standard Test Method for Steady-State Thermal Performance of Building Assemblies by Means of a Guarded Hot Box. 47. ASTM Standard C177-97, Standard Test Method for Steady-State Heat Flux Measurements and Thermal Transmission Properties by Means of the GuardedHot-Plate Apparatus. 48. S. Smith and C. Urso, Vuoto, 1999, 28, 1-2, 64. 49. N. E. Mathis, Measurement of Thermal Conductivity Anisotropy in Polymer Materials, Chemical Engineering Department, University of New Brunswick, Fredericton, NB, Canada, 1996, Ph.D. Thesis. 50. C. Dixon and N. Mathis, Presented at the Polyurethanes Expo ‘99, 1999, p.607. 51. J. K. Fremerey, Journal of Vacuum Science and Technology, 1985, A3, 3, 1715. 52. E. Kücükpinar, H. Güclü, A. S. Akkurt and F. Özkadi, Vuoto, 1999, 28, 1-2, 31. 53. E. Kücükpinar, F. Özkadi , A. Soyal and H. Güclü, Presented at the Polyurethanes Expo ‘99, 1999, p.339. 54. C. Boffito, M. Moraja and G. Pastore, Journal of Vacuum Science and Technology, 1997, 15, 4, 2391. 55. B. E. Lindenau and J. K. Fremerey, Journal of Vacuum Science and Technology, 1991, A9, 5, 2737.

211

Advances in Urethane Science and Technology 54. A. J. Hamilton, Vuoto, 1999, 28, 1-2, 27. 55. H. A. Fine and J. Lupinacci, Energy-Efficient Refrigerator Prototype Test Results, EPA-43D-R-94-011, June 1994. 56. E. A. Vineyard, J. R. Sand and R. H. Bohman, ASHRAE Transactions, 1995, 1, 1422. 57. E. A. Vineyard and J. R. Sand, presented at ORNL/CP-97450, Oak Ridge National Laboratory, TN, 1998. 58. G. J. Haworth, R. Srikanth and H. A. Fine, Presented at the 44th Annual International Appliance Technical Conference, Ohio, USA, 1993, p.1. 59. E. A. Vineyard, T. Stovall, K. E. Wilkes and K. W. Childs, ASHRAE Transactions, 1998, 104, 2. 60. T. F. Potter, D. K. Benson and L. K. Smith, Presented at ACEEE 1988, Proceedings from the Panel on Appliances and Equipment,1998, 4, 4.86.

212

5

Modelling the Stabilising Behaviour of Silicone Surfactants During the Processing of Polyurethane Foam: The Use of Thin Liquid Films Steven A. Snow, Udo C. Pernisz, Benjamin M. Nugent, Robert E. Stevens, Richard J. Braun and Shailesh Naire

5.1 Introduction The commercial production of polyurethane (PU) foam requires a tremendous diversity of ingredients. Common ingredients include isocyanates, polyether polyols, water, catalysts and stabilisers (surfactants). Most of the surfactants used in the foaming process are silicone surfactants, the major suppliers being Th. Goldschmidt AG, Witco and Air Products and Chemicals, Inc. Worldwide volume for silicone surfactants in polyurethane foam has been recently estimated at 30,000 tonnes/year [1]. In practice, a wide range of silicone surfactant structures are necessary to produce PU foam with a sufficiently broad range of physical properties. This structural diversity also reflects the need for the surfactant to perform many different tasks in foam formulations. The physical behaviour of these surfactants in the PU foam formulation process [2] has recently been reviewed. These roles of the surfactant include reducing surface (interfacial) tension [3-11], altering surface viscoelasticity [5, 7, 11, 12-15], and modifying polymer reaction kinetics and morphology [16-21]. The focus of this chapter is on the mechanism of PU foam stabilisation and how it is affected by silicone surfactants. In order to understand a complicated physical mechanism such as PU foam stabilisation, one must first be concerned with the thermodynamics and kinetics of the process. As a PU (or any) foam rises, energy is absorbed by the system and converted into a combination of excess gravitational and surface energy. For PU foams, this excess energy is created by both high-shear mixing and energy transfer from the chemical polymerisation reaction (between isocyanates and either polyols or water). One function of the silicone surfactant is to stabilise this foam until it is a self-supporting, gelatinous froth. This froth can further harden, yielding elastomeric or rigid foam products. Once hardened, the foam still contains excess gravitational and surface energy; however, the rate of dissipation of this energy is extremely low. Therefore, when discussing foam ‘stability’, one is specifically considering the rate of foam degradation. Foam degradation involves both physical and energetic changes in the foam. Four physical processes are commonly associated with PU foam degradation:

213

Advances in Urethane Science and Technology •

a decrease in foam volume,

•

a decrease in the number of bubbles (cells),

•

an increase in average cell size, and

•

a decrease in the dispersity of cell sizes.

These physical processes involve the release of excess gravitational and surface energy. Excess gravitational energy is released during the drainage of liquid from the foam. The liquid can drain from between the bubbles, thinning the liquid membrane known as the lamella. The liquid can also drain from the plateau borders in the foam. These borders are the relatively wide channels in the foam created at the intersection point of three lamellae. Mysels, Frankel and Shinoda [22] estimated that, in aqueous foams, the amount of gravitational energy released during degradation was five orders of magnitude less than the amount of surface energy released. It is assumed that this estimation roughly holds for the degradation of PU foams; therefore, the focus here will be on processes that reduce the surface energy of PU foams. Two processes, the diffusion of gas from small to larger bubbles (Ostwald Ripening) [23] and bubble coalescence, reduce the surface area of the foam and, therefore, the surface energy. Ostwald ripening involves the diffusive molecular transport of gas from small to large(r) bubbles. Ostwald ripening in foams leads to the counterintuitive phenomena of large bubbles in a foam expanding as smaller ones shrink. This process results both in a reduction in the number of bubbles and in the dispersity of bubble sizes. However, Owen [12] has demonstrated that Ostwald ripening plays only a minor role in PU foam stability; therefore, this chapter will focus on the process of bubble coalescence. Bubble coalescence involves the merging of two distinct bubbles into one, due to the rupture and recession of the lamella. Early in the PU foaming process, the rupture of a lamella at the edge of the PU foam will cause bubble collapse as the gas inside is lost to the surrounding atmosphere. This process will reduce the number of bubbles present and the foam volume. Lamella rupture within the interior of the foam will cause bubble coalescence, decreasing the number of bubbles, increasing the average bubble size, but not necessarily decreasing the foam volume. Later in the foaming process, as the PU gel network sets up, the lamella may rupture but the two bubbles can retain their separate identities. This process is known as cell opening. If many of the lamellae have ruptured in this fashion, the object is more correctly called a sponge instead of a foam. In this case, air can easily be pushed out of the sponge if it is loaded with weight (as for example when one sits on PU foam padding on a car seat). Therefore, the mechanical properties of the PU foam depend greatly on the degree of cell opening present. There is general agreement that a strong stabilising effect from the surfactant is necessary early in the

214

Modelling the Stabilising Behaviour of Silicone Surfactants… foam process. The stabilising role, if any, of the surfactant late in the foaming process remains a controversial issue [2, 19-21, 23-27]. This chapter will focus on the earlier stages of the foaming process, before the gelling of the liquid PU phase. The mechanism of rupture depends upon the thickness of the lamella film. Ultra-thin films, with thicknesses in the 1-20 nm range, are believed to rupture due to the amplification of thermally-generated surface waves [28-30]. However, studies by Artavia and Macosko [31] and Akabori and Fujimoto [32] demonstrated that most of the ruptured lamellae in a PU foam are in the 200-1000 nm thickness range. Therefore, it is assumed here that the mechanism of rupture of these thicker films is most important when considering PU foam. It has been extensively demonstrated that ‘thick’ liquid films rupture due to the formation (nucleation) of a pinhole in the film which spreads outward radially until the film is destroyed. The nucleation of the pinhole could result from either the presence of fluid instabilities in the film (for example, Marangoni, Rayleigh-Taylor or Benard instabilities) [33] or else by the presence of immiscible microphases which could serve as nucleation sites. For PU foams, it has been proposed that film rupture is nucleated by the phase separation of polyurea segments [19-21, 24-26]. It was postulated that the phase separation process could be retarded by the presence of silicone surfactants containing a high percentage of polyoxyethylene. The polyoxyethylene segment of the surfactant could interact strongly with polyurea via the formation of a hydrogen bonding network. However, in cases where the silicone contains lesser amounts of polyoxyethylene, such as in many commercial foam stabilisers, a direct effect of the concentration and structure of the silicone surfactant on this phase separation process has not been found [2, 26]. Circumstantial evidence of the influence of Marangoni instabilities on PU foam stability, and presumably film stability and the nucleation of pinholes, has been amply demonstrated [34]. Marangoni instability in a fluid film refers to a situation where a gradient of surface tension exists, stimulating compensating surface flows. For example, the stretching of the film reduces the surface concentration and, therefore, increases the surface tension. Furthermore, this stretched and thinned spot would be prone to rupture. However, within a certain range of surfactant concentration, this thin spot can re-thicken due to the diffusion/convection of surfactant and underlying bulk liquid to the thin spot (Marangoni flow). Once nucleated, the film hole can spread out radially at a rate inversely proportional to the bulk viscosity of the film. Spreading can occur until the film is destroyed. For example, in typical aqueous films, the rate of spreading is so high that high-speed photography is necessary to capture the event. However, in a gelling PU foam, the rate

215

Advances in Urethane Science and Technology of spreading can be very low, and, in fact, films can have multiple holes in them without being destroyed [24-26, 32]. Detailed investigations on the effect of silicone surfactants on the nucleation and growth of holes in PU films have not been reported. Furthermore, that is not the topic of this chapter which is specifically concerned with the effect of the silicone surfactant on the rate of drainage of PU films. However, this topic is highly related to film rupture as it has been shown that the probability of film rupture is an inverse squared function of the film thickness [35]. The time-dependent film thickness is a function of the drainage rate of the film.

5.2 Film Drainage Rate: Reynold’s Model and Further Modifications 5.2.1 Rigid Film Surfaces Thin film drainage has been investigated in detail both theoretically and experimentally [22, 28-30, 35]. An extensively applied physical model of a draining thin film is the one of Reynold’s, a form of which is shown in Equation 1: VRe = 2Fk3 3πμR 4

(1)

where VRe is the velocity of film thinning, F is the external force on the film (causing drainage), k is the film thickness, μ is the dynamic viscosity, and R is the film radius. The Reynold’s model assumes that: •

the liquid inside the film is Newtonian in its rheological behaviour.

•

the surfaces of the film are circular, immovable, parallel plates; surface velocity equals zero, surface viscosity is infinite.

•

the flow inside the film is laminar.

•

the flow inside the film is driven by gravity and/or pressure gradients and is resisted by the bulk viscosity of the liquid.

The dependencies of drainage (flow) rate on the variables of film thickness, area and bulk viscosity are quantified in Equation 1. For example, with decreasing film thickness k, the drainage rate decreases in proportion with k3; furthermore the drainage rate decreases with increasing viscosity.

216

Modelling the Stabilising Behaviour of Silicone Surfactants… Previous discussions [34] of PU film drainage rate (and ultimately foam stability and cell opening) focused on the dependence of rate on bulk viscosity. Per Reynold’s model, factors which increased the bulk viscosity, such as catalyst concentration and polyol reactivity, presumably decreased film drainage rate and yielded a more stable, less porous foam.

5.2.2 Mobile Film Surfaces The Reynold’s model and equation have also been modified to account for surface effects on film drainage rate [29]. To accomplish this, the assumption (boundary condition) of a surface velocity of zero must be relaxed. This change also decreases the surface viscosity from an infinite to a finite value. This process yields Equation 2. Allowing for a finite surface velocity increases the film drainage rate from that which would be expected under the Reynold’s conditions. V / VRe = 1 + 1 / e f

(2)

Where V/VRe is the measured velocity of thinning divided by the velocity of thinning under the conditions of the Reynold’s equation and ef is the surface mobility factor. The surface mobility factor is analysed in detail in [36]. This analysis demonstrates that two factors control the degree of surface mobility: the surface viscosity and the presence of surface tension gradients.

5.2.3 Surface Viscosity The rates of thinning of vertically-supported, thin liquid films of polyol solutions of various silicone surfactants have been measured [37]. It was found that the rate for films stabilised by a trimethylsilyl-capped polysilicate (TCP; a highly branched silicone not containing polyethers), was much lower than that for the films stabilised by common silicone polyether copolymer surfactants. The retardation of drainage rate was correlated with an increase in surface viscosity. Furthermore, it was noted that PU foams prepared using TCP were significantly more stable than those containing the commercial surfactants. Surface viscosity scales directly with the surfactant surface concentration, the intermolecular cohesion between the surface molecules and the intermolecular adhesion between the surfactant molecules and the underlying bulk liquid layer. Increasing the surface concentration decreases the average distance between molecules, therefore increasing the sum value of the attractive molecular forces between them. In

217

Advances in Urethane Science and Technology many cases, high surface concentration yields a low film drainage rate and a highly stabilised foam. The intermolecular cohesion is the result of the balance of attractive (van der Waals, polar bonding) and repulsive (ionic) intermolecular forces. The total cohesive energy is a product of the total area of cohesion and the cohesive energy per unit area. For example, the cohesion of the polyether chains of a typical silicone surfactant yields most of the surface viscosity contribution in PU films and foams [38]. Although it has not been investigated, one could speculate that the surface viscosity would scale with the length of the polyether chain. Increasing the chain length increases the potential area of contact between the chains. The unit area cohesive energy is a function of the strength of the intermolecular bonds present. For example, silicone surfactant molecules could possibly cohere via van der Waals bonding of the methyl groups of the siloxane backbone or else via polar bonding between the ethylene oxide units in the polyether chain. The van der Waals bonding between methyl groups is known to be weak; consequently, the surface viscosity of unsubstituted PDMS is close to zero [15, 38]. However, the polar bonding between the polyether chains would be much stronger and, therefore, the source of surface viscosity measured in silicone polyether copolymers. Intermolecular adhesion between the polyether chains of a typical silicone surfactant and the underlying ‘polyurethane’ liquid matrix could increase the surface viscosity contribution in PU films and foams. In order to maximise surface viscosity, and therefore minimise the drainage rate, the surfactant concentration, the intermolecular cohesion and adhesion should all be high. Later in the chapter evidence is presented correlating silicone surfactant concentration and structure (which influences the intermolecular cohesion and adhesion) to surface viscosity and film drainage rate.

5.2.4 Surface Tension Gradients An idealised model of the surface of a thin liquid film is one of a monolayer of evenlydistributed surfactant molecules. However, a more realistic model is one where the molecules are not evenly distributed; therefore, the surface concentration depends on surface position. The result of this heterogeneous distribution is that gradients of surface concentration, and therefore surface tension, are present. One example of this was pointed out in Section 5.1 on the effect of Marangoni instabilities on film rupture. Regarding film drainage, a surface tension gradient exerts a surface stress that can either impede or

218

Modelling the Stabilising Behaviour of Silicone Surfactants… accelerate the underlying bulk flow (drainage). In thin films, these gradients typically make the surface rigid, retarding flow and therefore decreasing, drainage rate. The gradients can be relieved by the diffusion of surfactant from areas of high to low concentration; the drainage rate of the film is proportionately increased. Diffusive surfactant fluxes are functions of the intensity of the concentration gradient and the diffusion coefficient of the surfactant. The intensity of the concentration gradient depends upon the overall surfactant concentration. At low or high concentrations, these gradients are relatively weak. At intermediate concentrations, they are quite strong. The diffusion coefficient of the surfactant is a function of its size and shape. The most simple and common case to analyse is where the surfactant assumes a spherical shape. In this case, the diffusion coefficient of the surfactant scales inversely with the solvated molecular volume (Stoke’s law). Overall, to accelerate film drainage, it is useful to have a high concentration of surfactant whose surface partition coefficient is also high. The surface partition coefficient is a measure of the tendency of the surfactant to adsorb at the surface instead of remaining in the bulk. In the case of silicone polyether surfactants, the coefficient scales with the ratio of silicone to polyether. Under these conditions (high surfactant concentration, high ratio of silicone to polyether) the flux of surfactant to the surface, relieving surface tension gradients, will also be high. A large diffusion coefficient (for the surfactant molecule, a small size and a compact shape) is also helpful. Later in this chapter correlations of silicone surfactant concentration, partition coefficient, and diffusion coefficient to film drainage rate will be discussed.

5.3 Experimental Investigation of Model, Thin Liquid Polyurethane Films and the Development of Qualitative and Semi-Quantitative Models of Film Drainage This section begins with a qualitative description of thin liquid PU films. This initial investigation had five goals in mind: to confirm that stable, vertically-oriented, thin liquid films could be prepared using mixtures of ingredients designed to model a PU foam, to study the hydrodynamic phenomena in the films, to compare the physical behaviour of these films to the behaviour of the more common aqueous soap films, to observe specific surfactant effects on the properties of these films, and to extrapolate conclusions about the behaviour of these films to operational PU foam. After this qualitative description, an accurate measurement of the drainage rates of these films was sought in order to study the effect of bulk and surface variables on the rate. In

219

Advances in Urethane Science and Technology order to make an accurate measurement, a novel interferometric method was developed and implemented. This method was validated by the discovery that film drainage rate scaled with the reciprocal of bulk viscosity as predicted by Reynold’s equation. Once this method was validated, the effect of silicone surfactant concentration and structure on the drainage rate of the films was investigated. Generally, the drainage rate displayed a maximum as a function of surfactant concentration. This maximum was consistent with a physical model where the two major influences on drainage rate are surface tension gradients and surface viscosity. As surfactant concentration increased, the retarding effect of gradients decreased, and the film drainage rate increased. However, above a certain concentration, the drainage retarding effect of surface viscosity overcame the gradient effect. The dependence of film drainage rate on silicone surfactant molecular structure was also systematically investigated. In order to understand this correlation, three physical parameters of the film affected by surfactant structure must be considered. These parameters are the surface partition coefficient, the surfactant molecular diffusion coefficient and the degree of intermolecular cohesion within the surface layer. Specifically, as the length of the polyether (solvophilic) portion of the surfactant increased, the surface partition coefficient decreased, the diffusion coefficient decreased, and the degree of cohesion increased. This resulted, at constant surfactant concentration, in a complex effect on the film drainage rate. A quantitative physical model of a draining vertical thin film was developed from first principles. The starting point was the Navier-Stokes equation. The initial model featured a fixed-surface, wedge-shaped vertical film, with immobile surfaces. This is essentially the Reynold’s model modified to this film shape and orientation. Good agreement of the predictions of this model with experimental data was obtained. The next model relaxed the condition of fixed film shape (allowed for curvature in the film surface) and analysed the effects of the menisci on the film drainage. Analysis of the Navier-Stokes equation was simplified by the application of the lubrication approximation. The results from this analysis agreed extremely well with experimental values both in terms of film drainage rate and the changes in film shape with time. The development and growth of bulges and waves on the bottom of the film were particularly intriguing. This phenomenon was experimentally observed in these films. Finally, models were developed where the condition of infinite surface viscosity was relaxed. This allowed the analysis of surfactant effects on film drainage, in particular surface viscosity and surface transport. Specifically, the model predicted the decrease in drainage rate as surface viscosity increases, as expected from the qualitative models and measured experimentally. The effect of surface transport was significantly less than that of surface viscosity.

220

Modelling the Stabilising Behaviour of Silicone Surfactants…

5.3.1 Experimental Details 5.3.1.1 Film Formulation Unless noted otherwise, the experimental work in this chapter involved a model flexible slabstock PU foam formulation [39] at the instant of mixing (see below): •

20.00 g VORANOL 3137 (3100 MW polyether copolymer, The Dow Chemical Company),

•

5.46 g toluene,

•

0.20 g DABCO DC 198 (surfactant Air Products and Chemicals, Inc.)

For simplicity’s sake, chemical reactions were avoided. This was accomplished by substituting toluene for toluene diisocyanate (TDI) in the formulation. This formulation had a nominal viscosity of about 0.8 Pa-s.

5.3.1.2 Description of the Experimental Interferometer The following experimental variables were addressed during the design and construction of the interferometer: •

control of film area,

•

control of the withdrawal rate of the film,

•

control of spurious vibration potentially leading to film rupture,

•

control of solvent evaporation,

•

precise vertical alignment of the film,

•

optical control of the interferometric measurement (illumination with plane parallel light, use of monochromatic light, uniform illumination of the film surface),

•

Imaging of the film including magnification, presentation on a video screen, and video recording capabilities, and

•

Precise measurement of film lifetimes.

221

Advances in Urethane Science and Technology Consideration of these points led to the experimental setup sketched in Figure 5.1. The film is formed between two narrow blades set by 10 mm apart and supported in a surrounding frame; a film of 20 mm height could be formed between the blades. The frame was vertically clamped on an optical bench. A glass cuvette with rectangular cross section was fabricated with a closely fitting lid through which the clamp rod fits tightly in order to prevent evaporation of the solvent and at the same time to allow the cuvette to be lowered and raised to adjust the bulk liquid level or to draw a film. The cuvette stands on a small pedestal which was driven by a computer-controlled stepper motor assembly allowing film size control at velocities of the receding or advancing bulk liquid level up to 25 mm/s. The sample holder assembly was mounted on an optical bench. The two nearly plane-parallel surfaces of the film constitute a Fizeau interferometer which produces interference fringes in the reflected light [40]. The fringes are lines of constant film thickness. Film drainage data were acquired by generating monochromatic light by means of an interference filter (λ = 505 nm) or with a HeNe laser (λ = 632.8 nm).

Figure 5.1 Schematic of the thin liquid film interferometer.

222

Modelling the Stabilising Behaviour of Silicone Surfactants…

Figure 5.2 Depiction of film formation. The film is formed within the boundaries of the inner frame by the lowering of the liquid level at a controlled rate. The frame and liquid are within the transparent walls of the enclosed glass cuvette.

5.3.1.3 Film Formation The process of film formation is depicted in Figure 5.2. Films were formed by lowering the sample liquid from a point where the vertical frame was completely immersed to a point where it was only partially immersed. In most experiments, the vertical film so formed was left in direct contact with the bulk liquid. Films were withdrawn from the bulk liquid at a rate varying from 0.25 to 50 mm/s.

5.3.2 Qualitative Description of Polyurethane Films In a typical experiment, stable, vertical, liquid films were formed from the model PU formulation (see Section 5.3.1.1). These films were stable for two to five minutes. The series of photographs in Figure 5.3 depicts many of the physical features of these films. Thirty to sixty seconds after film formation, dark horizontal interference fringes were observed that initially appeared at the top of the film and steadily moved downward. Over time, the number of fringes decreased. Flow patterns, including eddy currents, fingering patterns and swirls, appeared at the bottom and sides of the films and rose upwards. The rate and amount of these flows decreased as the lifetime of the film increased. Films were subjected to a rapid raising and lowering of the bulk liquid level. These motions strongly stimulated the surface flows previously mentioned.

223

Advances in Urethane Science and Technology

Figure 5.3 The visual sequence of a draining film. The first photograph in the sequence is in the top left hand corner and the sequence of photographs proceeds clockwise. Photograph 2: one minute after formation. The dark horizontal lines faintly present are the interference fringes. Photograph 4: 3-5 minutes after film formation. Fewer fringes are present and span the height of the film. The reduction in the number of fringes versus Photograph 2 is consistent with a significant amount of film drainage. Photograph 3: 8-12 minutes after film formation. Photograph 1: approximately 15 minutes after film formation.

The pattern of horizontal fringes observed on the face of the films suggested a vertical wedge shape, with the film thickness increasing as one descends the film. The wedge shape of the vertical PU films has a number of implications for operational PU foam. Any cell window in PU foam does not have uniform thickness. This non-uniformity has been reported in at least two studies [24-26, 32]. In addition, it would be expected that the rupture of the cell window would occur at its thinnest point [24-26, 32]. As seen in Figure 5.3, the interference fringes progress down the face of the film and the distance between the fringes increases as the film continues to drain. This indicates a decrease in the film thickness gradient during the drainage of the film. The drainage of the wedge-shaped film can be visualised in that the sides of the wedge, intersecting at the

224

Modelling the Stabilising Behaviour of Silicone Surfactants… top of the film, pivot inwards as fluid drains out of the bottom. This is a collapsing inwards of the wedge. As the film is draining, interference fringes radiate outwards from the top centre of the film (photographs 2-4 in Figure 5.3). This radiation suggests that both horizontal and vertical flow processes are present. These flow processes are driven by both gravity (vertical flow) and suction into the Plateau borders of the film (horizontal flow). Once the fluid has entered the Plateau border, it flows down a relatively wide channel into the bulk liquid. The Plateau borders in these films can be seen in Figure 5.3. They are the dark regions between the film and the frame. The width of the Plateau channel increases as one descends the face of the film, giving it roughly a triangular shape. Plateau borders are also present in operational polyurethane foams. In fact, horizontally-aligned films within these foams can only drain due to Plateau border suction. The individual Plateau borders within the PU foam form a network which allows for liquid drainage down the foam. Ultimately, this network becomes the structural ‘struts’ of the foam. As reported in the literature, interference fringes in many vertical aqueous films are horizontal with little curvature. These films also displayed ‘mobile’ surfaces, rapid and extensive surface flows and rapid rates of drainage [22, 28-30]. These features were all linked to the presence of a low surface viscosity. Based on our experimental observations, the PU films stabilised by DABCO DC 198 have low surface viscosities. This conclusion is supported by reports of measurements of low surface viscosities in polyol solutions containing silicone polyether surfactants [12, 15, 38]. The phenomena of edge turbulence, fingering and upward flows in aqueous soap films have been extensively investigated [22, 28-30, 41-46]. These hydrodynamic phenomena have been defined as marginal regeneration and gravity convection. Marginal regeneration refers to a process where, simultaneously, thick films are sucked into the Plateau border and thinner films are pulled out. This exchange results in a net increase of material in the Plateau border, essentially draining the film. Within the film, after the material exchange, the new thin spots then migrate upward (gravity convection) until they reach a height where their thickness equals the film thickness. Instead of describing these phenomena in PU films as ‘gravity convection ‘, it is proposed that this flow near the surface of the film is driven by surface tension gradients. These are examples of Marangoni flows, and would proceed from an area of high surface tension to one of low surface tension. Marangoni flows are extremely common and one well-known example is the phenomenon of ‘tears’ in a glass of port wine or brandy. This view is also gaining some acceptance in the field of aqueous thin liquid films. Stein [4244] recently proposed that the flows observed in mobile-surfaced, aqueous soap films are Marangoni flows.

225

Advances in Urethane Science and Technology The surface tension gradients necessary to drive this type of flow can arise from three different physical phenomena: surface tension gradients are necessary to support the weight of the vertical film, flows in the film or the channel of the Plateau border impose stresses on the surface monolayer which are balanced by a surface tension gradient, and extension or compression of the surface layer can create transient surface tension gradients. Regarding the cell windows of operational PU foam, it is apparent that surface tension gradients are constantly present, due to all three mechanisms discussed previously. In particular, the constant stretching of the bubbles during the process of foam growth acts as a potent stimuli for these flows.

5.3.3 Quantitative Measurement of Film Drainage Rates: Bulk and Surface Effects 5.3.3.1 A Method to Measure Film Drainage Rates: The ‘Collapsing Wedge’ Model The distribution of the interference fringes in a liquid film can be viewed as a contour map of the film thickness, with each successive fringe representing a section of constant film thickness (isopach). Therefore, the distance between fringes throughout the film gives the gradient of film thickness. In order to precisely measure PU film drainage rates, a physical model correlating the time change in the fringe density (ds/dt; the quantity that can be measured in the laboratory) to the time change in volume (dV/dt, the film drainage rate) of the film must be applied. Qualitatively it appears that the drainage of the film collapsed the sides of the film wedge inwards. During this collapsing process, the two sides of the wedge act as if they were hinged together at the top of the film. With this geometric model of film shape, the drainage rate of the film, dV/dt, can be expressed mathematically by Equation 3. dV c ds = WL2 dt 2 dt

(3)

Where V is the film volume, t is time, c is the change in film thickness per interference fringe, W and L are the film width and height, respectively, and s is the fringe density. If both sides of Equation 3 are divided by the cross-sectional area (WcsL) of the film at the bottom Equation 4 is obtained:

226

Modelling the Stabilising Behaviour of Silicone Surfactants… (1/WScL) dV/dt = L/2s ds/dt

(4)

The quantity on the left of Equation 4 is the flux. If it is assumed that the flux is a hyperbolic function of time, namely, (5)

(1/WScL) dV/dt = b/t + t0 where b is a drainage constant and t0 is the time where hyperbolic drainage begins. Solving the differential equation for s yields Equation 6. s = k0 (t + t0 )

m

with m = −

2b L

(6)

where k0 is the fringe density at t = t0 As Equation 6 shows, a log-log plot of fringe density versus time should be linear for t > t0 with a slope of m (90

Mechanical Stability, minutes

30-850

30-140

75-770

110-1,200

-

-

-

120-420

37-55 75-130

Particle size distribution

Properties of DPUR

75.1

56.2

215.5

211.1

-

-

-

161.8

58.3

nm

Average Particle size

-98.8

-60.6

-4.5

-7.9

-

-

-

-21.2

-54.6

nm

Zeta potential mV

+50

MFFT °C

nt

nt

+98.0

-88.0

+123.1

+58. 0

Tg °C

Table 6.4 Properties of ASD used as starting materials for synthesis of MDPUR according to method 3

Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…

285

Advances in Urethane Science and Technology In Figures 6.14, 6.15 and 6.16 the particle size distribution, zeta potential and TEM micrograph of dispersion particles of typical ASD from Table 6.3 (ASD sample number 15) are presented.

Figure 6.14 Particle size distribution for typical ASD synthesised in this study (ASD 15 from Table 6.3).

286

Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…

Figure 6.15 Zeta potential for typical ASD synthesised in this study (ASD 15 from Table 6.3).

Figure 6.16 TEM micrograph of dispersion particles of ASD (see Table 6.3). Reproduced with permission from Professor I. A. Grickova at Lomonosov University, Moscow. 287

Advances in Urethane Science and Technology

6.5.2 Synthesis of MDPUR and MDPUR-ASD Hybrid polyurethane-urea-acrylic/styrene polymer dispersions were prepared according to methods ‘1a’, ‘1b’, ‘2’ and ‘3’ described in Section 6.3.2. Dispersions designated as MDPUR-ASD were made by polymerisation of monomers in DPU according to the methods 1a, 1b and 2 while dispersions designated as MDPUR were made by synthesis of DPUR in ASD according to method 3. In all syntheses the ratio of polyurethane-urea to acrylic/styrene polymer in the hybrid was 2:1. Tables 6.5, 6.6, 6.7 and 6.8 show the compositions of MDPUR-ASD prepared according to methods 1a, 1b, 2 and 3, respectively.

Table 6.5 Composition of MDPUR-ASD prepared according to method 1a (polymerisation of monomers in DPUR with continuous feeding of the monomers) Designation of MDPUR-ASD

Designation of starting DPUR (see Table 6.1)

Monomer

Coalescent (NMP) content in MDPURASD, %

MDPUR-ASD 22

DPUR 240

MM

0

MDPUR-ASD 23

DPUR 244

BA

0

MDPUR-ASD 24

DPUR 244

BA, MM, S

0

MDPUR-ASD 14

DPUR 244

S

0

MDPUR-ASD 259

DPUR 345

BA, MM, S

7

MDPUR-ASD 268

DPUR 346

BA, MM, S

2

MDPUR-ASD 274

DPUR 346

BA, MM, S

2

MDPUR-ASD 300

DPUR 354

MM, S

0

MDPUR-ASD 307

DPUR 356

BA, MM, S

3.3

MDPUR-ASD 309

DPUR 356

BA, MM, S

3.3

Table 6.6 Composition of MDPUR-ASD prepared according to method 1b (polymerisation of monomers in DPUR after swelling DPUR particles with monomers) Designation of starting DPUR (see Table 6.1)

Monomer

Coalescent (NMP) content in MDPURASD, %

MDPUR-ASD 270

DPUR-346

BA, MM, S

2

MDPUR-ASD 272

DPUR-346

BA, MM, S

2

MDPUR-ASD 304

DPUR-356

BA, MM, S

3.2

Designation of MDPUR-ASD

288

PPUR 288

PPUR 289

PPUR 333

MDPUR-ASD 98

MDPUR-ASD 102

MDPUR-ASD 198 PTMG-2000 Irpurate

PTMG-2000

PTMG-2000

PTMG-2000

Polyol

BA,MM,S

BA,MM,S

BA,MM,S

BA,MM,S

Monomer

-

3

3

3

Coalescent (NMP) content in MDPUR-ASD %

2.75

3.66

2.48

3.96

2.68

2.68

2.50 2.50

Pract.

Theoret.

NCO, %

Nt

730

760

760

Viscosity at 25 °C, mPa-s

Properties of starting prepolymer-ionomer

PTMG-2000

PTMG-2000

PTMG-2000

PTMG-2000

MDPUR 245

MDPUR 247

MDPUR 250

MDPUR 347

PTMG-2000

PTMG-2000

MDPUR 243

MDPUR 348

Polyol

Designation of MDPUR

3.6

3.6

11.6

11.6

11.6

11.6

Coalescent (NMP) Content in MDPUR, %

ASD 266

ASD 264

ASD 20

ASD 18

ASD 17

ASD 15

Designation of starting ASD

3.12

3.12

2.74

2.74

2.74

2.74

theoret.

NCO, %

2.79

2.85

2.51

2.24

2.34

2.64

Pract.

5000

5700

4000

4200

4500

4000

Viscosity at 25 °C, mPa-s

Properties of starting prepolymer-ionomer

Table 6.8 Compositions of MDPUR prepared according to method 3 (synthesis of DPUR in ASD). Properties of starting prepolymer-ionomers are also included

Nt – not tested

PPUR 288

MDPUR-ASD 97

Designation of MDPUR-ASD

Designation of starting prepolymerionomer (PPUR)

Table 6.7 Composition of MDPUR-ASD prepared according to method 2 (diluting of the prepolymer-ionomer in monomers, polymerisation and crosslinking of prepolymer-ionomer). Properties of starting prepolymerionomers are also included. Synthesis and Characterisation of Aqueous Hybrid Polyurethane-Urea…

289

Advances in Urethane Science and Technology

6.5.3 Investigation of the Effect of Various Factors on the Properties of Hybrid Dispersions 6.5.3.1 Method of Hybrid Dispersion Synthesis and Kind of Initiating System The results of investigations of the effect of method of hybrid dispersion synthesis (1a, 1b, 2 or 3 – see Section 3.2) on the properties of dispersions as well as of films and coatings made from them are presented in Tables 6.9 to 6.11 (dispersions prepared using water-soluble initiator) and in Tables 6.12 to 6.14 (dispersions prepared using ‘redox’ initiating system). In all dispersions the chemical structure of the polyurethane-urea and acrylic/styrene polymer component was the same (see relevant tables in Section 6.5.2). All the dispersions contained a similar low level (2-3.6%) of NMP. Particle size distribution and zeta potential of a typical hybrid dispersion obtained using a water-soluble initiator are presented in Figures 6.17 and 6.18, respectively.

290

34.2 31.4 41.6

MDPURASD 270

MDPURASD 97

MDPUR 348

1b

2

3 200

10

10

9

Viscosity MPa-s

7.4-7.7

6.8-7.1

6.8-7.1

7.1-7.4

pH

15

15

15

15

Mechanical Stability Min

Nt

0.7/0.02/ 0.015

Nt

Nt

Free monomer, BA/MM/S, %

55-520

50-120

50-670

30-230

Particle size distribution

Properties of dispersions

115.4

108.4

9 4. 4

61. 9

Average part size nm

-23.0

-60.7

-39.1

-18.4

Zeta pot MV